Abstract IA12: An in vivo high throughput shRNA screening platform for identifying ways to target genomic alterations

Comprehensive characterization of genomic alterations prevalent in head and neck squamous cell carcinoma (HNSCC) has been performed by our group and others, and more recently by The Cancer Genome Atlas (TCGA) project. A significant barrier towards targeting the genomic alterations collectively found in HNSCC is that the genomic landscape of this disease is dominated by mutations to tumor suppressor genes that are inherently difficult to target, with very few mutated oncogenes present. Although potentially targetable oncogenic drivers may exist within chromosomal regions of copy number gain, the large number of genes within these regions makes identifying the precise targets challenging. To overcome some of these difficulties, we have been leveraging high throughput loss of function shRNA screens performed in vivo in order to identify co-dependencies and potential ways to target genomic alterations in HNSCC. Using a large panel of genomically characterized established HNSCC cell lines in conjunction with several different shRNA libraries, including ones that target druggable genes (i.e. FDAome), genes involved in DNA repair, and candidate HNSCC driver genes, we hope to link shRNA targets that are vital for in vivo growth and survival to specific genomic subtypes present within HNSCC. Following pooled in vitro infection with these libraries, HNSCC cell lines are injected into the flanks of mice where in vivo selection takes place, and dropout of barcoded shRNAs targeting genes advantageous for tumor growth is subsequently analyzed from harvested tumors by next generation sequencing. The FDAome library contains shRNAs to nearly 200 genes immediately targetable by FDA approved drugs, allowing identification of druggable co-dependent pathways through simple lethality. By comparing dropout of shRNAs from tumors of mice treated with or without carboplatin, genes that chemosensitize tumors in vivo are also being identified within the FDAome and DNA repair shRNA libraries. Screens performed with our custom HNSCC shRNAs library targeting many genes from regions recurrently amplified in large subsets of patient tumors should identify which genes within these regions are true drivers and potentially interesting drug candidates. Thus far, we have completed in vivo screens employing the FDAome and DNA damage libraries on four different HNSCC cell lines, comprised of 3 different NOTCH1 mutants and one NOTCH1 wild type cell line, which collectively include two HPV-positiv