Tumor diameters were measured with digital calipers, and tumor vol- ume was calculated by the following formula: tumor volume (mm3) = shorter diameter2 × longer diameter/2. The tumor volume data are pre- sented as means ± SD (n = 15). Our study was approved by the Animal Care and Use Committee of the Renji Hospital affiliated find more to Shanghai Jiao Tong University School of Medicine. All animal procedures were performed according to the guidelines developed by the China Council on Animal Care and

the protocol approved by the Renji Hospital affiliated to Shanghai Jiao Tong University School of Medicine. The genomic sequencing and survival data analyzed in this study were obtained from The Cancer Genome Atlas (TCGA) ZD1839 mouse GBM data set [16].

The published versions of these data sets include 586 cancer cases with sequencing data and clinical information. The corresponding reverse phase protein array (RPPA) data of TCGA GBM data set were obtained using the cBioPortal online data portal (Memorial Sloan-Kettering Cancer Center, New York, NY) [17], which include quantified expression of 122 proteins and 43 phosphoproteins involved in various signaling pathways. Exam- ples of such pathways include p53 signaling, retinoblastoma (RB), AMP-activated protein kinase (AMPK), PTEN, phosphatidylinositol 3-kinase (PI3K)/protein kinase B (PKB), receptor tyrosine kinase (RTK)/RAS GTPase, and epidermal growth factor receptor signaling, and other sequences. A complete list of antibodies in the protein micro- array can be accessed from TCGA data portal (http://tcga-data.nci.nih. gov/tcga/, Memorial Sloan-Kettering Cancer Center). The patients with upper or lower quarter

Pten protein expression were determined according to the levels detected by RPPA (respectively ranked as 25% highest or 25% lowest). We obtained the sensitivity profiles of 59 human brain tumor 2-hydroxyphytanoyl-CoA lyase cell lines to 131 anticancer drugs from the Cancer Cell Line Encyclo- pedia (CCLE; Broad Institute, Cambridge, MA) database [18]. The half-maximal inhibitory concentration (IC50) was used as a measure of the effectiveness of a drug on the cell lines. The mutation spectrum of TP53 in these cell lines was similar with that in the TCGA data sets. Survival analysis was carried out in R program using the “survival” package as described previously [19]. In the Kaplan-Meier (log-rank) survival test and Cox regression models, the censored status is in- dicated when the patient was still alive (or cancer free) at the time of follow-up. The Cox regression model included cancer type as a covariant, and the P value for mutation type is calculated after adjustment for the factor of cancer type. The hazard ratios (HRs) and 95% confidence intervals (CIs) were also determined for each mutation. The effects of different p53 mutations were compared to nonsense mutations as an indication of gain-of-function (GOF) effect.