Abstract
Background
Ependymomas represent the third most common malignant pediatric brain tumor accounting for up to 10% of CNS malignancies in childhood. Recent studies have identified novel molecular subgroups within this histologically defined tumor entity. Supratentorial ependymomas harbor either YAP or RELA gene fusions whereas in the PF compartment PFA and PFB can be discriminated by presence of hypermethylation. The standard of care comprise surgery followed by radiotherapy, while chemotherapy seems to be widely ineffective. Despite intensive treatment, the most frequent ependymoma subtypes - ST-RELA and PFA - frequently recur, highlighting the need for predictive biomarkers in pediatric ependymoma. Recently, FGFR1 and FGFR3 protein levels have been found elevated in aggressive ependymoma and were associated with worse outcome. Accordingly our study follows the hypothesis whether FGFRs represent a novel promising target in pediatric ependymoma.
Material and Methods
To this end we investigated mRNA levels of FGFR3 and its splice variants FGFR3B and FGFR3C by qPCR in 44 ependymoma tissue samples and compared them to an in silico dataset. The effects of the FGFR inhibitors nintedanib and ponatinib on cell viability and on intracellular signaling were investigated in several ependymoma cell models of different molecular subgroups by MTT-assays and Western blot, respectively. We further tested whether the expression of a dominant negative (dn) FGFR1 or FGFR3 impacts cell survival. Finally, the response to nintedanib was evaluated in a RELA fusion-positive ependymoma xenograft model.
Results
Analyses of FGFR mRNA levels in an ependymoma online database exhibited FGFR3 to be highly expressed and point to the highest levels in ST-RELA. We already analyzed mRNA levels of FGFR3 and its splice variants in 44 tissue samples of our ependymoma cohort (n=54). Expression levels were significantly increased in ST and highest in ST-RELA as compared to PF tumors. Regarding alternative splicing, the FGFR3c variant was predominantly detected in our collection. Several of our ependymoma cell culture models exhibited sensitivity towards nintedanib and ponatinib in the low micromolar range. Our results further highlight an impact of FGFR inhibition on MAP-Kinase and PI3K-signaling in two cell models (VBT73 and VBT145) from consecutive recurrent ST-RELA ependymoma. Interestingly, and in contrast to the expression of a dnFGFR3, dnFGFR1 exhibited no effect on cell survival. Finally, treatment of ST-RELA xenografts (VBT145) in immunodeficient mice with nintedanib significantly extended survival as compared to vehicle controls.
Conclusion
Summarizing, our findings confirm that FGFR3 plays a role in pediatric ependymoma biology and appears a promising target for several ependymoma subgroups.