While tumour cells exhibited very strong FUBP1 protein expression levels, weaker FUBP1 staining Sirolimus was observed in both CD31-positive endothelial cells (Figure 5E) and NeuN-positive neurones (data not shown). As it has been suggested from sequence analyses that all FUBP1 mutations identified in oligodendrogliomas may lead to FUBP1 protein truncation, we examined whether the FUBP1 protein expression analysis can be used as a convenient screening parameter to detect FUBP1 mutations [1]. For this purpose, we screened 15 glioma patients with oligodendroglial

differentiation (six cases with absence of FUBP1 protein expression on tumour cells and nine showing moderate or high FUBP1 levels also in glioma cell nuclei) by sequencing all FUBP1 exons (excluding exon 6 due to technical reasons). The results from the mutation screen are presented in Table 2. FUBP1 immunohistochemistry was able to predict FUBP1 mutations with a sensitivity of 100% and a specificity of 90%. With this approach, we were able to identify a novel nonsense mutation (p.Q508X), which was found in WHO grade III oligodendroglioma lacking FUBP1 protein expression (Figure 6). This novel mutation was predicted to inactivate the

encoded protein due to the creation of a stop codon. FUBP1-negative cases were significantly associated with 1p/19q LOH (P = 0.0027) and showed a trend for IDH1 mutation

(R132H) (P = 0.0953) in gliomas with oligodendroglial differentiation. In addition, the constant C59 wnt molecular weight preservation of nuclear FUBP1 expression in neurones, microglia, reactive astrocytes and endothelial cells in the otherwise FUBP1-negative tumour samples suggests that the identified genetic alterations are somatic and not germline Interleukin-2 receptor mutations thereby serving as internal positive control. Here we report on the FUBP1 expression profile of human gliomas and its association with established diagnostic markers including mutated IDH1 (R132H), MIB-1 index (Ki-67) as well as genetic alterations including 1p/19q LOH and its relation to the FUBP1 mutation status. In normal brain tissue, strong FUBP1 protein expression was only observed in neuronal cells (Figure S2). These findings correlate with previous reports showing that FUBP1 potentially contributes to the neuronal differentiation of human embryonic stem cells and interacts with SMN in the foetal and adult mouse brain, thereby suggesting that it also contributes to neuronal cell survival [8,10]. In contrast to the selective neuronal expression pattern observed in the normal CNS tissues, FUBP1 expression levels are increased in all glioma subtypes independent of the subtype, both at mRNA (Figure S3) and at protein levels (Figures 1-3).