Fig 2: Binding of FGFR2 peptide to Barrett's neoplasiaOn representative images collected with confocal microscopy of human esophageal specimens ex vivo, SRR*-Cy5.5 (red) shows minimal staining to (A) squamous (SQ) and (B) Barrett's esophagus (BE) and strong binding (arrows) to (C) high-grade dysplasia (HGD) and (D) esophageal adenocarcinoma (EAC). (E-H) Anti-FGFR2 antibody labeled with AF488 (green) was used as a positive control, and shows weak staining to SQ and BE but strong binding (arrows) to HGD and EAC. We quantified the fluorescence intensities from the mean of a set of 3 boxes with dimensions of 30×30 μm2 placed over cells, shown in panels (C) and (G). From n = 28, 33, 22, and 17 specimens of SQ, BE, HGD, and EAC, respectively, we found significantly greater mean fluorescence intensity from HGD and EAC compared with that for BE with (I) SRR*-Cy5.5 and (J) AF488-labeled anti-FGFR2 using an ANOVA model with terms for 4 means on log-transformed data. (K) ROC curve shows 87% sensitivity and 70% specificity for detecting Barrett's neoplasia (HGD and EAC) at a T/B ratio of 3.0. (L-O) Merged images shows co-localization of peptide (red) and antibody (green) binding. We determined a Pearson's correlation coefficient of ρ = 0.59, 0.54, 0.52 and 0.59 for SQ, BE, HGD and EAC, respectively. Representative histology (H&E) are shown for (P) SQ, (Q) BE, (R) HGD, and (S) EAC.

Fig 3: Effect of miR-628 on the expression level of FGFR2 in PCa cells. The relative expression level of (A) miR-628 and (B) FGFR2 was determined in the LNCaP cell line following transfection with miR-628-5p mimic. The relative expression level of (C) miR-628 and (D) FGFR2 was determined in the LNCaP cell line following transfection with miR-628-5p inhibitor. The protein expression level of FGFR2 was determined by western blot analysis in LNCaP cells following transfection with (E) miR-628-5p mimic or (F) miR-628-5p inhibitor. **P<0.01 vs. control. miR, microRNA; FGFR2, fibroblast growth factor receptor 2; PCa, prostate cancer; LNCaP, human prostatic adenocarcinoma cell line.

Fig 4: AZD4547 exerts tumor suppressive role by inhibiting FGFR2 and YAP1 signaling.a The IC50 value of AZD4547 in AGS and MGC-803 cells. b Monoclonal colony formation of MGC-803 cells under the treatment of rhFGF18 (50 ng/ml), AZD4547 (5 µM), or AZD4547/rhFGF18, respectively (**p < 0.001). c FGFR and the MEK–ERK signaling were activated by the rhFGF18 stimulation (50 ng/ml) but quenched by AZD4547 (5 µM). d The intensity of migration related F-actin (green) and cell division-related α-Tubulin (red) was analyzed in MGC-803 cells by immunofluorescence staining (scale bar, 20 μm). In the statistical bar chart, 1 to 4 represent vehicle, rhFGF18, AZD4547, and rhFGF18 + AZD4547, respectively. e AZD4547 dose-dependently decreased the downstream targets of YAP1, including CTGF and c-Myc (24 h). f rhFGF18 (50 ng/ml) partly restored CTGF and c-Myc expression in AZD4547 (25 µM for AGS, 5 µM for MGC-803, 25 µM for MKN28) treated cell lines.

Fig 5: Validation of specific FGFR2 peptide binding to cellsOn confocal microscopy, we observed strong binding of SRR*-Cy5.5 (red) to surface of QhTERT cells that express (A) FGFR2b and (B) FGFR2c compared with (C) wild type. (D-F) Minimal signal is seen with the scrambled peptide SPS*-Cy5.5. (G-I) Strong binding is seen with anti-FGFR2 antibody labeled with AF488 (green) used as positive control. All experiments were performed in triplicate. (J) Quantified results show significantly higher mean fluorescence intensities for SRR*-Cy5.5 versus SPS*-Cy5.5 (control). We log-transformed and averaged measurements for 3 random cells on each of 3 slides per condition, and fit an ANOVA model with terms for 6 means. (K) Western blot shows protein expression level of FGFR2 for each cell.