Acquiring resistance to the FGFR inhibitor may

Acquiring resistance to the FGFR inhibitor may determine treatment outcomes in patients with GC dependent on FGFR2 amplification, and we investigated the mechanism of AZD4547 resistance in FGFR2-amplified GC cells. We established a FGFR2-amplified GC cell line resistant to the selective FGFR inhibitor, AZD4547. We observed elevated phosphorylation of EphB3 and a pattern of expression levels of various proteins characteristic of EMT. Treatment with an EphB3 TKI abrogated the expression pattern, suggesting the EMT-mediated resistance to AZD4547 was due to activation of EphB3. Signaling through EphB3 was observed to be transmitted to mTOR via the Ras-ERK1/2 pathway, supporting use of mTOR inhibitors as one of therapeutic strategies to overcome resistance to the selective FGFR inhibitor.
Material and methods
Results
Discussion Accumulating evidence of the role of FGF/FGFR signaling in oncogenesis of several cancers has led to the clinical development of agents targeting activated FGFR. AZD4547 was developed as a selective FGFR inhibitor to complement the relatively weak activity of multi-kinase inhibitors, such as dovitinib and ponatinib (Dieci et al., 2013). Despite its potent activity, anti-cancer effects of AZD4547 has been hampered by emerging resistance. Therefore, we generated a GC cell line resistant to AZD4547 to elucidate the mechanism of resistance. Activation of alternative RTK signaling pathways, one of commonly suggested mechanisms for acquisition of resistance to targeted agents, was demonstrated using the receptor phosphorylation assay in our study. We observed that tyrosine phosphorylation of EphB3 was elevated in SNU-16R cells. Recent works on EphB3 have reported its controversial roles in cancer progression versus tumor suppression. EphB3 as a tumor suppressor was presented in colorectal cancer at first. Batlle et al. reported the downregulation of mRNA levels of EphB3 in tumor tissues of colon cancer. The in vivo study presented here also showed the loss of EphB3 Gadodiamide sale in mice accelerated colorectal tumorigenesis (Batlle et al., 2005). Another study also proved tumor suppressive function of EphB3 in colon cancer by showing reduced expression of EphB3 in the clinical cohort with advanced Dukes’ stage, and inhibited tumor growth in EphB3-overexpressing colon cancer cells and an in vivo xenograft model (Chiu et al., 2009). Furthermore, Li, et al. reported tumor suppressive function of EphB3 in NSCLC, as ligand-dependent phosphorylation of EphB3 inhibited cell migration. On the other hand, kinase-independent overexpression of EphB3 in NSCLC cells promoted tumorigenesis by stimulating cell migration and growth, suggesting RTK phosphorylation-dependent effects on cancer (Ji et al., 2011; Li et al., 2012). In our study, upregulated phosphorylation of EphB3 was demonstrated in SNU-16R cells, and inhibition of its tyrosine phosphorylation resulted in inhibition of tumor cell growth and invasiveness. A previous study reported kinase-activated EphB3 suppressed metastatic capability of NSCLC by formation a PP2 A/RACK1/Akt complex leading to decreased Akt phosphorylation [29]. Our study also showed reduced phosphorylation of Akt in SNU-16R cells, consistent to the previous report, however, phosphorylation of mTOR and its downstream substrates was observed to be elevated. This discrepant finding led us to examine activation status of the Ras/Raf/MEK1/2/EKR1/2 pathway. Increased phosphorylation of proteins involved in this pathway and their suppression by EphB3 TKI and siRNA in SNU-16R cells made us assume that mTOR was activated through the Ras-ERK1/2 pathway (Fig. 6). As a matter of fact, several studies have reported cross-activation of mTOR signaling by the Ras-ERK1/2 pathway by means of regulating phosphorylation of tuberous sclerosis complex (TSC) 2, mTOR complex 1, and 4E-BP1, supporting our assumption (Carriere et al., 2011; Mendoza et al., 2011; Roux et al., 2004; She et al., 2010). Taken together, the potent tumorigenic effect of kinase-activated EphB3 shown in this study was considered to be mediated by activation of the Ras/Raf/MEK1/2/ERK1/2 signaling cascade.