Mutations in PIK3CA are infrequent in neuroblastoma

BACKGROUND: Neuroblastoma is a frequently lethal pediatric cancer in which MYCN genomic amplification is highly correlated with aggressive disease. Deregulated MYC genes require co-operative lesions to foster tumourigenesis and both direct and indirect evidence support activated Ras signaling for this purpose in many cancers. Yet Ras genes and Braf, while often activated in cancer cells, are infrequent targets for activation in neuroblastoma. Recently, the Ras effector PIK3CA was shown to be activated in diverse human cancers. We therefore assessed PIK3CA for mutation in human neuroblastomas, as well as in neuroblastomas arising in transgenic mice with MYCN overexpressed in neural-crest tissues. In this murine model we additionally surveyed for Ras family and Braf mutations as these have not been previously reported. METHODS: Sixty-nine human neuroblastomas (42 primary tumors and 27 cell lines) were sequenced for PIK3CA activating mutations within the C2, helical and kinase domain "hot spots" where 80% of mutations cluster. Constitutional DNA was sequenced in cases with confirmed alterations to assess for germline or somatic acquisition. Additionally, Ras family members (Hras1, Kras2 and Nras) and the downstream effectors Pik3ca and Braf, were sequenced from twenty-five neuroblastomas arising in neuroblastoma-prone transgenic mice. RESULTS: We identified mutations in the PIK3CA gene in 2 of 69 human neuroblastomas (2.9%). Neither mutation (R524M and E982D) has been studied to date for effects on lipid kinase activity. Though both occurred in tumors with MYCN amplification the overall rate of PIK3CA mutations in MYCN amplified and single-copy tumors did not differ appreciably (2 of 31 versus 0 of 38, respectively). Further, no activating mutations were identified in a survey of Ras signal transduction genes (including Hras1, Kras2, Nras, Pik3ca, or Braf genes) in twenty-five neuroblastic tumors arising in the MYCN-initiated transgenic mouse model. CONCLUSION: These data suggest that activating mutations in the Ras/Raf-MAPK/PI3K signaling cascades occur infrequently in neuroblastoma. Further, despite compelling evidence for MYC and RAS cooperation in vitro and in vivo to promote tumourigenesis, activation of RAS signal transduction does not constitute a preferred secondary pathway in neuroblastomas with MYCN deregulation in either human tumors or murine models.