Thus far, we have been in the process of performing experiments in preparation for the screen. We have performed array CGH on a panel of glioma stem-like cell lines that were obtained from Dr. Cameron Brennan (TS600, TS603, TS616, MSKCC) or Dr. Phil Tofilon (GBAM1, NCI ROB), and an immortalized normal human astrocyte line which will be used as a control. We determined that two glioma stem cell lines have a single copy of chromosome 10 (TS600 and TS616), one has a single copy deletion of 10q (TS543), and three cell lines have primarily intact chromosome 10 (TS603, GBAM1). There are 15 kinases on chromosome 10 representing 15 kinase families. We have also performed extensive optimization of conditions under which the stem-like cells will be infected with the murine stem cell virus (MSCV) hairpin library. We have established a viral purification and infection method that reduces stem-like cell toxicity and differentiation, while optimizing transduction efficiency. Validating chromosome 10 kinase status in GBM stem-like cell lines. We will first measure the gene expression of the 16 protein kinases on chromosome 10 in the deleted and intact cells in Table 2 by real time quantitative PCR. We will then compare these results to the array CGH data in Figure 2 to validate that single copy loss of chromosome 10 coincides with reduced expression levels of the deleted kinases. Synthetic lethal kinase screen in chromosome 10-deleted glioma stem-like cells. We will perform a pooled shRNA screen with a validated library of hairpins directed against the entire kinase family in collaboration with Dr. Ji Luo of the NCI Medical Oncology Branch. Dr. Luo has extensive experience with large-scale genetic screens and will provide the library and assistance with the design of the experiment. The glioma stem-like cells will each be infected with the kinase hairpin pool in triplicate at an MOI of 1-2 and an average representation of 1,000 infected cells per shRNA. Virus will be removed 24 hrs later, and a fraction of the cells will be collected at 48 post-infection and saved for later sequencing as the initial samples (PD=0). Puromycin-selected cells will then be continuously passaged until final harvesting at PD=12. Genomic DNA will be extracted from the PD=0 and PD=12 cells, and shRNAs will be PCR-amplified and sequenced with Solexa sequencing by the NIH CCR Cancer Genetics Branch sequencing facility. A false-discovery rate of 20% as well as a 2-fold shRNA decrease cutoff will be applied to the dataset for each cell line to identify shRNAs which reduce cell viability in the screen. Validation of candidate kinases for on-target effects. Kinase shRNAs absent from the chromosome 10 deleted cell lines will be further analyzed to determine whether their absence is due to synthetic lethality or cell death due to off-target effects of the hairpin. Candidate kinases will be inhibited individually with commercially-available siRNAs or pharmacological agents and cell viability will be compared between chromosome 10 intact and deleted cells. Identification of the critical redundant kinases lost on chromosome 10. To determine the identity of the lost kinases that are redundant with the positive hits in the shRNA screen, we will add back each kinase in Table 2 to the chromosome 10 deleted cells and assess viability upon knockdown of the candidate kinase. Validating the synthetically lethal kinase as a therapeutic target in chromosome 10-deleted tumors. We will enlist the candidate kinases in multiple in vitro and in vivo assays to determine its relevance as a potential therapeutic target. We will perform validation assays for cell proliferation, apoptosis, invasion, anoikis, and soft agar growth that are routinely performed in our laboratory. We will also determine the feasibility of therapeutically targeting the candidate kinases in cancers other than GBM by performing these assays in additional tumor lines with frequent deletions of chromosome 10, such as prostate and melanoma. An advantage of using the glioma stem-like cells in this screen is that they generate orthotopic xenograft tumors in immunocompromised mice that are indistinguishable histologically from human gliomas. Therefore, to determine the relevance of the synthetically lethal kinases as drug targets, we will measure tumor formation and overall survival in mice implanted with glioma stem-like cells in which the kinases are inhibited with shRNAs or small molecule inhibitors, and compare these outcomes between the chromosome 10-deleted and intact cells.
