Introduction: Pilocytic astrocytoma (PA) is the most common childhood brain tumor. This tumor can occur throughout the central nervous system, with roughly 50% of cases arising outside of the cerebellum. Tumors in these non-cerebellar locations are often difficult to treat surgically, leading to multiple tumor recurrences. PA can therefore become a chronic disease, with patients experiencing substantial morbidities. Alterations in the MAPK pathway, particularly BRAF, have previously been identified in ∼80% of PAs. Interestingly, however, the majority of those cases without a recognized change are non-cerebellar.

Methods: To investigate the full range of genetic alterations occurring in PA, we used Illumina HiSeq technologies to perform whole-genome sequencing of matched tumor and germline DNA from 47 patients, with corresponding RNA sequencing data for 35 tumors.

Results: The average somatic mutation rate in PA was extremely low, at 0.065/Mb, with an average of only 1.8 non-synonymous coding single nucleotide variants (SNVs) per tumor - almost ten times lower than we have recently reported for medulloblastoma. We found several novel alterations in known PA-related genes, including two new oncogenic BRAF fusions. Most strikingly, however, we identified mutations at two hotspots in the FGFR1 receptor tyrosine kinase in 4/6 centrally located PAs lacking any other MAPK pathway change. Two of these cases also carried a mutation in a downstream adaptor protein, PTPN11 (Shp2). Interestingly, germline mutations of PTPN11 are associated with Noonan syndrome (NS), and there are case reports of NS patients developing PAs. Screening of additional non-cerebellar PAs revealed four further cases with an FGFR1 mutation. All PAs, regardless of MAPK pathway alteration, displayed highly elevated expression of FGF2, indicating a general role for ligand-mediated activation of the FGFR1/MAPK pathway in PA tumorigenesis. Notably, the same FGFR1 mutations were also identified in four midline pediatric glioblastomas (GBM), a highly malignant brain tumor, suggesting a possible common origin for a subset of these two entities, despite their dramatically different clinical course.

Conclusion: Altogether, MAPK alterations were identified in 96% of PAs, with very few other changes, confirming the concept of PA as a single-pathway disease. Our results also suggest that a subset of centrally located, FGFR1-driven pediatric PAs and GBMs may share common origins. Most importantly, they reveal a novel therapeutic target in clinically relevant subsets of childhood glioma.

Citation Format: David TW Jones, Barbara Hutter, Natalie Jäger, Andrey Korshunov, Marcel Kool, Sally R. Lambert, Dong Anh Khuong Quang, Adam M. Fontebasso, Marina Ryzhova, Hans-Jörg Warnatz, Thomas Zichner, Jan O. Korbel, Stephan Wolf, Marie-Laure Yaspo, Keith L. Ligon, Mark W. Kieran, Benedikt Brors, Jörg Felsberg, Guido Reifenberger, V. Peter Collins, Nada Jabado, Roland Eils, Peter Lichter, Stefan M. Pfister, ICGC PedBrain Tumor Project. Recurrent FGFR1 hotspot mutations represent a novel therapeutic target in childhood astrocytoma. [abstract]. In: Proceedings of the 104th Annual Meeting of the American Association for Cancer Research; 2013 Apr 6-10; Washington, DC. Philadelphia (PA): AACR; Cancer Res 2013;73(8 Suppl):Abstract nr 4594. doi:10.1158/1538-7445.AM2013-4594