Purpose: Farnesylthiosalicylic acid (FTS) is a Ras inhibitor that dislodges all active Ras isoforms from the membrane.We assessed the ability of FTS to reverse the transformed phenotype of neurofibromatosis type 1 (NF1)^associated tumor cell lines of malignant peripheral nerve sheath tumor (MPNST). Experimental Design: nf1 mutations were genotyped, allelic losses were analyzed, and neurofibromin expression levels were determined in MPNST cell lines ST88-14, S265P21, and 90-8. The effects of FTS on GTP-bound Ras (Ras-GTP) and its prominent downstream targets, as well as on cell morphology, anchorage-dependent and anchorage-independent growth, and tumor growth in mice, were assessed. Results: The MPNSTcell lines were biallelic, NF1inactive, and neurofibromin deficient. We show that FTS treatment shortened the relatively long duration of Ras activation and signaling to extracellular signal-regulated kinase, Akt, and RalA in all NF1-deficient MPNST cell lines (NF1 cells) to that observed in a non-NF1, normally expressing neurofibromin MPNSTcell line. These effects of FTS led to lower steady-state levels of Ras-GTP and its activated targets. Both anchoragedependent and anchorage-independent growth of NF1cells were dose dependently inhibited by FTS, and the inhibition correlated positively with Ras-GTP levels. NF1cells were found to possess strong actin stress fibers, and this phenotype was also corrected by FTS. NF1tumor growth in a nude mouse model was inhibited by oral FTS. Conclusions: FTS treatment of NF1cells normalized Ras-GTP levels, resulting in reversal of the transformed phenotype and inhibition of tumor growth. FTS may therefore be considered as a potential drug for the treatment of NF1.

Ionizing radiation is the most established risk factor for meningioma formation. Our aim was to evaluate the main effect of selected candidate genes on the development of meningioma and their possible interaction with ionizing radiation in the causation of this tumor. The total study population included 440 cases and controls: 150 meningioma patients who were irradiated for tinea capitis in childhood, 129 individuals who were similarly irradiated but did not develop meningioma, 69 meningioma patients with no previous history of irradiation, and 92 asymptomatic population controls. DNA from peripheral blood samples was genotyped for single nucleotide polymorphisms (SNP) in 12 genes: NF2, XRCC1, XRCC3, XRCC5, ERCC2, Ki-ras, p16, cyclin D1, PTEN, E-cadherin, TGFB1, and TGFBR2. SNP analysis was done using the MassArray system (Sequenom, San Diego, CA) and computerized analysis by Spectro-TYPER. Logistic regressions were applied to evaluate main effect of each gene on meningioma formation and interaction between gene and radiation. Intragenic SNPs in the Ki-ras and ERCC2 genes were associated with meningioma risk (odds ratio, 1.76; 95% confidence interval, 1.07-2.92 and odds ratio, 1.68; 95% confidence interval, 1.00-2.84, respectively). A significant interaction was found between radiation and cyclin D1 and p16 SNPs (P for interaction = 0.005 and 0.057, respectively). Our findings suggest that Ki-ras and ERCC2 SNPs are possible markers for meningioma formation, whereas cyclin D1 and p16 SNPs may be markers of genes that have an inverse effect on the risk to develop meningioma in irradiated and nonirradiated populations.

While genetic factors clearly play a key role in colorectal cancer (CRC) pathogenesis and in determining its phenotypic features, the precise genes that involved are largely unknown. To gain insight into these genes, consecutive Israeli CRC patients were genotyped using SNPs from within candidate genes: APC, ␤-Catenin, K-RAS, DCC, P16, PTEN, RB1, P15, APOE, ERCC2, P53, MTHFR and hMSH2. Genotyping of consecutive, unselected colorectal cancer patients was done mostly by utilizing the MassAR-RAY technology (Sequenom) and to a lesser extent DGGE, ARMS and direct DNA sequencing. Correlation of genotypes with specific phenotypic features was carried out for all patients and separately for the Ashkenazim. Overall, 456 patients were analyzed, the majority (64.25%) being of Ashkenazi origin; mean age at diagnosis was 65.6 ؎ 14 (range 25-90 years), and the mean follow-up was 4.7 ؎ 0.28 (range 0 -30 years). Statistically significant associations were noted between SNPs in ␤-catenin and APOE and a positive family history of cancer (␤-catenin: p‫,430.0؍‬ APOE: p‫;)330.0؍‬ tumor location and a DCC SNP (p‫)830.0؍‬ and the P53 R72P mutation and survival (p‫.)6330.0؍‬ In Ashkenazi patients, ERCC2 and MTHFR genes' SNPs were associated with age at diagnosis (ERCC2: p‫,520.0؍‬ MTHFR: p‫;)5000.0؍‬ a P53 polymorphism, APOE and Rb SNPs with a family history of cancer (P53 p‫430.0؍‬ ;APOE p‫,40.0؍‬ Rb p‫؍‬ 0.022); DCC SNP with tumor location (p‫;)410.0؍‬ and p15 SNP with tumor grade (p‫.)230.0؍‬ This preliminary study shows that genetic factors play a role in determining CRC phenotypic features and that a larger cohort with longer follow-up is clearly needed.Key words: colorectal cancer; SNP; genotype-phenotype correlation; Hardy Weinberg equilibrium; tumor phenotype Colorectal cancer (CRC) is the second leading cause of cancer morbidity and cancer death among adult Americans, with 135,000 newly diagnosed cases and 57,000 deaths in 2001. 1 In Israel, there are about 3,300 new cases each year. 2,3 Similar to other cancer types, colon cancer has a multifactorial etiology, where environmental and genetic factors converge to result in the tumorous phenotype. Genetic factors play a role in familial colorectal cancer in the form of highly penetrant low prevalent genes and in sporadic cases where low penetrance high prevalence genes are seemingly involved in the pathogenesis. 4 Moreover, genetic factors also play a pivotal role in determining colorectal cancer phenotypic features: tumor stage, grade, location within the colon and age at diagnosis, as well as putatively controlling therapeutic response and survival parameters. 5 The precise genes, molecular mechanisms and pathways that determine these phenotypic features and therapeutic responses are largely unknown, but several candidate genes exist. These genes are candidates as they have been shown to be somatically involved in CRC (e.g., p53, DCC and APC), involved in syndromic familial CRC predisposition (e.g., hMSH2 and PTEN), participate in molecular pathways implicated in colonic epit...