Unraveling the Mechanisms of Sensitivity to Anti-FGF Therapies in Imatinib-Resistant Gastrointestinal Stromal Tumors (GIST) Lacking Secondary KIT Mutations

We showed previously that inhibition of KIT signaling in GISTs activates FGFR-signaling pathway rendering cancer cells resistant to receptor tyrosine kinase inhibitor (RTKi) imatinib mesylate (IM) (Gleevec) despite of absence of secondary KIT mutations and thereby illustrating a rationale for the combined (e.g., KIT- and FGFR-targeted) therapies. We show here that long-term culture of IM-resistant GISTs (GIST-R1) with IM substantially down-regulates KIT expression and induces activation of the FGFR-signaling cascade, evidenced by increased expression of total and phosphorylated forms of FGFR1 and 2, FGF-2, and FRS-2, the well-known adaptor protein of the FGF-signaling cascade. This resulted in activation of both AKT- and MAPK-signaling pathways shown on mRNA and protein levels, and rendered cancer cells highly sensitive to pan-FGFR-inhibitors (BGJ 398, AZD 4547, and TAS-120). Indeed, we observed a significant decrease of IC50 values for BGJ 398 in the GIST subclone (GIST-R2) derived from GIST-R1 cells continuously treated with IM for up to 12 months. An increased sensitivity of GIST-R2 cells to FGFR inhibition was also revealed on the xenograft models, illustrating a substantial (>70%) decrease in tumor size in BGJ 398-treated animals when treated with this pan-FGFR inhibitor. Similarly, an increased intra-tumoral apoptosis as detected by immunohistochemical (IHC)-staining for cleaved caspase-3 on day 5 of the treatment was found. As expected, both BGJ 398 and IM used alone lacked the pro-apoptotic and growth-inhibitory activities on GIST-R1 xenografts, thereby revealing their resistance to these TKis when used alone. Important, the knockdown of FGFR2, and, in much less content, FGF-2, abrogated BGJ 398's activity against GIST-R2 cells both in vitro and in vivo, thereby illustrating the FGF-2/FGFR2-signaling axis in IM-resistant GISTs as a primary molecular target for this RTKi. Collectively, our data illustrates that continuous inhibition of KIT signaling in IM-resistant GISTs lacking secondary KIT mutations induced clonal heterogeneity of GISTs and resulted in accumulation of cancer cells with overexpressed FGF-2 and FGFR1/2, thereby leading to activation of FGFR-signaling. This in turn rendered these cells extremely sensitive to the pan-FGFR inhibitors used in combination with IM, or even alone, and suggests a rationale to re-evaluate the effectiveness of FGFR-inhibitors in order to improve the second-line therapeutic strategies for selected subgroups of GIST patients (e.g., IM-resistant GISTs lacking secondary KIT mutations and exhibiting the activation of the FGFR-signaling pathway).

Establishment and Characterization of Multi-Drug Resistant p53-Negative Osteosarcoma SaOS-2 Subline

AIM

To establish a p53-negative osteosarcoma (OS) SaOS-2 cellular subline exhibiting resistance to specific chemotherapeutic agents, including topoisomerase II inhibitors, taxanes, and vinca alkaloids.

METHODS

The OS subline exhibiting resistance to the chemotherapeutic agents indicated above was generated by the stepwise treatment of the parental SaOS-2 cell line with increasing concentrations of doxorubicin (Dox) for 5 months. Half-inhibitory concentrations (IC50) for Dox, vinblastine (Vin), and paclitaxel (PTX) were calculated by a colorimetric MTS-based assay. Crystal violet staining was used to assess cellular viability, whereas the proliferation capacities of cancer cells were monitored in real-time by the i-Celligence system. Expression of apoptotic markers (e.g., cleaved PARP and caspase-3), DNA repair proteins (e.g., ATM, DNA-PK, Nbs1, Rad51, MSH2, etc.), and certain ABC transporters (P-glycoprotein, MRP1, ABCG2, etc.) was assessed by western blotting and real-time PCR. Flow cytometry was used to examine the fluorescence intensity of Dox and ABC-transporter substrates (e.g., Calcein AM and CMFDA) and to assess their excretion to define the activity of specific ABC-transporters. To confirm OS resistance to Dox in vivo, xenograft experiments were performed.

RESULTS

An OS subline generated by a stepwise treatment of the parental SaOS-2 cell line with increasing concentrations of Dox resulted in an increase in the IC50 for Dox, Vin, and PTX (~6-, 4-, and 30-fold, respectively). The acquisition of chemoresistance in vitro was also evidenced by the lack of apoptotic markers (e.g., cleaved PARP and caspase-3) in resistant OS cells treated with the chemotherapeutic agents indicated above. The development of the multidrug resistance (MDR) phenotype in this OS subline was due to the overexpression of ABCB1 (i.e., P-glycoprotein) and ABCC1 (i.e., multidrug resistance protein-1, MRP-1), which was evidenced on both mRNA and protein levels. Due to increased expression of MDR-related proteins, resistant OS exhibited an excessive efflux of Dox. Moreover, decreased accumulation of calcein AM, a well-known fluorescent substrate for both ABCB1 and ABCC1, was observed for resistant OS cells compared to their parental SaOS-2 cell line. Importantly, tariquidar and cyclosporin, well-known ABC inhibitors, retained the intensity of Dox-induced fluorescence in resistant SAOS-2 cells. Furthermore, in addition to the increased efflux of the chemotherapeutic agents from Dox-resistant OS cells, we found higher expression of several DNA repair proteins (e.g., Rad51 recombinase, Mre11, and Nbs1, activated forms of ATM, DNA-PK, Chk1, and Chk2, etc.), contributing to the chemoresistance due to the excessive DNA repair. Lastly, the in vivo study indicated that Dox has no impact on the SaOS-2 Dox-R xenograft tumor growth in a nude mouse model.

CONCLUSIONS

An acquired resistance of OS to the chemotherapeutic agents might be due to the several mechanisms undergoing simultaneously on the single-cell level. This reveals the complexity of the mechanisms involved in the secondary resistance of OS to chemotherapies.