Discovery of a small molecule ligand of FRS2 that inhibits invasion and tumor growth

Abstract

Purpose: Aberrant activation of the fibroblast growth factor receptor (FGFR) family of receptor tyrosine kinases drives oncogenic signaling through its proximal adaptor protein FRS2. Precise disruption of this disease-causing signal transmission in metastatic cancers could stall tumor growth and progression. The purpose of this study was to identify a small molecule ligand of FRS2 to interrupt oncogenic signal transmission from activated FGFRs.

Methods: We used pharmacophore-based computational screening to identify potential small molecule ligands of the PTB domain of FRS2, which couples FRS2 to FGFRs. We confirmed PTB domain binding of molecules identified with biophysical binding assays and validated compound activity in cell-based functional assays in vitro and in an ovarian cancer model in vivo. We used thermal proteome profiling to identify potential off-targets of the lead compound.

Results: We describe a small molecule ligand of the PTB domain of FRS2 that prevents FRS2 activation and interrupts FGFR signaling. This PTB-domain ligand displays on-target activity in cells and stalls FGFR-dependent matrix invasion in various cancer models. The small molecule ligand is detectable in the serum of mice at the effective concentration for prolonged time and reduces growth of the ovarian cancer model in vivo. Using thermal proteome profiling, we furthermore identified potential off-targets of the lead compound that will guide further compound refinement and drug development.

Conclusions: Our results illustrate a phenotype-guided drug discovery strategy that identified a novel mechanism to repress FGFR-driven invasiveness and growth in human cancers. The here identified bioactive leads targeting FGF signaling and cell dissemination provide a novel structural basis for further development as a tumor agnostic strategy to repress FGFR- and FRS2-driven tumors.

Keywords: Bioactive small molecule compound; Cell invasion; FGFR; FRS2; Protein–protein interaction interference; Thermal proteome profiling.

Fig. 1

Identification and validation of small molecule compound ligands of the FRS2-PTB domain. a) Workflow of experimental procedures. b) Quantification of representative spheroid invasion assay (SIA) in DAOY cells treated with indicated compounds. Violin plot with median and quartiles of distances of invasion at 10 µM and adjusted P value from n = 3 technical replicas at 1, 5 and 10 µM compound concentrations are shown. Red dotted line: Maximal repression of bFGF-induced invasion, green dotted line: Maximal bFGF-induced invasion. c) nanoDSF analysis and compound-induced ΔT_m of FRS2_PTB of two independent measurements are shown. * marks compounds with self-fluorescence exceeding the protein’s fluorescence. Para: paracetamol. d) K_D of shortlisted compounds binding to GB1-FRS2_PTB determined by measuring change in initial fluorescence using MST trace analysis. Fitting with a signal–noise ratio < 5 is considered no binding. e) EC50 of shortlisted compounds for the inhibition of bFGF-induced collagen I invasion in DAOY cells. Log(Y) transformed and normalized invasion distances and corresponding SD of n = 3 technical replicas are plotted against compound concentrations. f) nanoDSF analysis of GB1-FRS2_PTB binding of 3.18 analogs. Mean and SD of compound-induced ΔT_m at 10 µM compound concentration is shown. n = 3 technical replicas. g) nanoDSF analysis of FRS2_PTB binding of shortlisted hits and bioisoesters of 3.18 and 7. Mean and SD of compound-induced ΔT_m is shown of n = 3 technical replicas. h) Quantification of SIA in DAOY cells with bioisosteres of compounds 3.18 (11) and 7 (12, 13, 14) at 1, 5 and 10 µM compound concentrations. Mean and SD from n = 3 biological replicas are shown

Fig. 2

NMR screening of compounds 3.14 and 13. Reference spectrum with assignments of 3.14 a) and 13 b) (upper trace) and WaterLOGSY (middle) and STD (bottom) of 25 µM GB1-FRS2-PTB and 500 µM 3.14 or 13 at 280 K. Resonances exhibiting protein-interaction in the STD and/or the WaterLOGSY spectra are highlighted in light green, while resonances that might be obscured by buffer signals are indicated by light orange strips. (c and d) Cartoon of secondary structure showing the backbone trace of FRS2_PTB with side-chains of the binding pocket depicted as sticks. Compounds 13 (c and e) and 3.14 (d and f) are depicted as purple and blue sticks. (d and f) Protein surface within a 8 Å sphere around the ligand atoms. Residues with significant CSPs are highlighted in orange. Unassigned residues are shown in dark green. Note that, additional WaterLOGSY signals with a negative signal phase and originating from the arginine of the buffer system were detected for 3.14. These arginine resonances also interfered with the resonances a,b and c of compound 13, rendering the corresponding WaterLOGSY and STD signals ambiguous

Fig. 3

Compound-mediated blockade of collagen I invasion in human cancer cell models. a) SIA analysis with shortlisted compounds in different human cancer cell models. Mean distance or area (RT-112) of invasion and SD of n = 3 experiments at 10 µM compound concentrations are shown. Dotted lines: Blue: Basal mean invasion unstimulated, green: maximal mean invasion bFGF-stimulated, red: Maximal inhibition of invasion by BGJ398, orange: 50% of maximal bFGF-induced mean invasion. b) Heat map of adjusted P values of SIA analysis shown in a and in S4a for DAOY cells. c) Heat map of percent change of invasion relative to unstimulated control of SIA shown in a and in S4a for DAOY cells. d) Boyden chamber transwell migration assay. Mean and SD of n = 4 technical replicas and one-way ANOVA adjusted P value statistics are shown (* p =  < 0.05, *** p =  < 0.001). e) SIA of RT-112 cells comparing inhibitory effect of compound with BGJ398. f) SIA of RT-112 cells transfected with either two different FRS2-specific or their corresponding control siRNAs. Violin plot with median and quartiles of distances of invasion area of invasion of ≥ 6 spheroids and one-way ANOVA adjusted P value statistics from representative experiment are shown in e) and f) (**** p =  < 0.0001)

Fig. 4

Compound-mediated blockade of MAPK pathway activation in human cancer cell models. a) IB analysis of bFGF-induced FRS2_Y436 phosphorylation in DAOY cells. b) IB analysis of FRS2_Y436, ERK1/2_{Thr202/Tyr204} and AKT_S476 phosphorylation in compound treated DAOY cells stimulated with bFGF. c) Bar diagrams depicting the quantification of the phosphorylation of FRS2_Y436 and ERK1/2_{Thr202/Tyr204} relative to unstimulated control. d) Left: IB analysis of ERK1/2_{Thr202/Tyr204} phosphorylation in compound treated SW780 cells stimulated with bFGF. Right: Bar diagram depicting the quantification of ERK1/2_{Thr202/Tyr204} phosphorylation relative to unstimulated control. e) Left: IB analysis of ERK1/2_{Thr202/Tyr204} and AKT_S476 phosphorylation in compound treated M059K cells stimulated with bFGF. Right: Bar diagram depicting the quantification of ERK1/2_{Thr202/Tyr204} phosphorylation relative to unstimulated controls. Bar diagrams in c-e show mean fold change of phosphorylation of n = 3 independent experiments, SD and one-way ANOVA adjusted P values of comparison to SFM + bFGF are shown

Fig. 5

Bioavailability of compounds and in vivo efficacy of compound 7. a) Plasma exposure levels after 10 mg/kg peroral (PO, compounds E12 and 7) or 1 mg/kg intravenous (IV) (compounds E12, 3.14 and 7) administration in mice quantified at intervals between 0.167 and 24 h. Means and SD of n = 3 animals are shown. b) EC50 curves of SIA with compound 7 in different human cancer cell models in the presence of 100 ng/ml bFGF. c) IC50 curves of CellTiter Glo cell viability analysis with compound 7 in different human cancer cell models in full growth medium. Log(Y) transformed and normalized invasion distances (b) or viability (c) and corresponding SD of n = 3 technical replicas are plotted against compound concentrations. d) Tumor volumes of SKOV3 flank tumors of mice treated with DMSO, 10 mg/kg BGJ398 or 200 mg/kg compound 7 relative to volumes at start of treatment. A two-way ANOVA analysis with Turkey’s multiple comparison test was performed (* p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001). Lower: Tumor growth curves of the individual tumors. e) Upper: Tumor volumes of AGS flank tumors of mice treated with DMSO, 10 mg/kg BGJ398 or 200 mg/kg compound 7 relative to volumes at start of treatment. Lower: Tumor growth curves of individual tumors. f) IHC analysis of pFRS2_Y436 staining in tumor samples of the SKOV3 flank tumor model at d74 (end of treatment)

Fig. 6

Level of FRS2 determines compound 7 efficacy. a) IB analysis of the expression levels of total endogenous FRS2 in DAOY wt cells in comparison to FRS2 expression in DAOY FRS2-FLAG cells. b) IB analysis of the expression and FRS_Y436 phosphorylation levels of total endogenous FRS2 in DAOY siCTL cells in comparison to DAOY siFRS2 cells 48 h post transfection. Cells were either maintained in SFM or in SFM supplemented with 100 ng/ml bFGF for five min. c) Validation by IB analysis of compound 7 activity towards bFGF-induced FRS2_Y436 phosphorylation in DAOY cells. d) Comparison of compound 7 repression of bFGF-induced collagen I invasion in DAOY wt and DAOY FRS2-FLAG cells by SIA. Truncated violin plot with median and quartiles of distances of invasion at increasing compound 7 concentrations is shown. A 2-way ANOVA test with Šídák's multiple comparison was performed (** p < 0.01 and **** p < 0.0001). e) Comparison of EC50 of collagen I invasion for compound 7 in DAOY wt and DAOY FRS2-FLAG cells. Log(Y) transformed and normalized invasion distances and corresponding SEM of n = 10 technical replicas plotted against compound concentrations are shown. f) Dose effect on cumulated invasion distances in DAOY wt and DAOY FRS2-FLAG cells. g) Comparison of compound 7 repression of bFGF-induced collagen I invasion in DAOY siCTL and DAOY siFRS2 cells. Violin plot with median and quartiles of distances of invasion at increasing compound 7 concentrations is shown. A 2-way ANOVA test with Šídák's multiple comparison was performed (** p < 0.01, *** p < 0.001 and **** p < 0.0001). h) Comparison of EC50 of collagen I invasion for compound 7 in DAOY siCTL and DAOY siFRS2 cells. Log(Y) transformed and normalized invasion distances and corresponding SEM of n = 10 technical replicas are plotted against compound concentrations. i) Dose effect on cumulated invasion distances in DAOY siCTL and DAOY siFRS2 cells. Means and SD of > 5 technical replicas are shown

Fig. 7

On/off target activity of compound 7 in cells and lysates. a) IB of FRS2 and tubulin from whole cell CETSA after DMSO or compound 7 treatments. b) Nonlinear fit of FRS2 and tubulin abundance in CETSA from DMSO- or compound 7-treated whole cells shown in a. c) Left: Pie chart of percentage stabilized and destabilized proteins (p < 0.05) from whole cell TPP analysis. Right: Volcano plot of ∆T_m from proteins of which high quality melting curves were obtained in both conditions. Red dots ∆T_m ≥ -5 °C, green dots: ∆T_m ≥ 5 °C. d) Upper: IB of FRS2 and tubulin in cell lysate CETSA after DMSO or compound treatment. Lower: Nonlinear fit of FRS2 and tubulin abundance of CETSA from DMSO or 7-treated cell lysates. e) Scatterplot dot plot of T_m distributions of all proteins with high quality melting curves of whole cell and lysate TPP analyses. f) Upper: Pie chart of percentage stabilized and destabilized proteins (p < 0.05) from lysate TPP analysis. Lower: Volcano plot of ∆T_m from proteins of which high quality melting curves were obtained in both conditions. Red dots ∆T_m ≥ -5 °C, green dots: ∆T_m ≥ 5 °C. g) Bar plot with gene symbols of proteins with significantly altered T_m (p < 0.05) after manual inspection of melting curves. Green bars: ∆T_m ≥ 5 °C