Abstract C057: Delineating tissue-specific oncogene selection with autochthonous CRISPR activation screening

Abstract

Abstract Background: While the genetic landscapes of pancreatic ductal adenocarcinoma (PDAC) and non-small cell lung cancer (NSCLC) are well described, the contribution of individual genes and transcriptomic programs to tumor initiation, progression and metastasis remain incompletely understood. To quantify the oncogenic potential of individual genes and identify transcriptomic programs that drive progression and metastasis, we have developed a platform for in vivo CRISPR activation (CRISPRa) screening that allows us to delineate tissue-specific oncogene selection. Material and Methods: Our platform consists of a conditional, CRISPRa-competent, murine tumor model (PPKS; P53(F/F), Kras(LSL-G12D/+), R26(LSL-SAM), coupled with targeted delivery of Cre-recombinase/sgRNA library encoding lentivirus [1]. We designed a CRISPRa screening library targeting 452 of the most commonly amplified and mutated genes across human adenocarcinoma. Using this library, we generated pancreatic and lung tumors, and harvested tissues and metastatic lesions at multiple time points (6 days, 4 weeks, 8 weeks, and endpoint). The change in guide distribution over time was then determined using NGS and bioinformatic analysis. To confirm the oncogenic potential of the most highly enriched genes, single gene-activated tumors were compared to non-targeted controls. Results: We found that essentially all guides (99.9%) were represented in vivo at the 6-day time point. The degree of guide selection increased over time and was significantly higher in the pancreas relative to the lung, indicating a higher degree of parallel tumor evolution in the lung. Lung samples were characterized by Egfr, Sox2 and Myc displaying gradually increasing enrichment over time and Egfr displayed the highest level of enrichment in endpoint lung tumors. In stark contrast, pancreas samples were almost completely dominated by Myc activation at the 4-week timepoint, with increased enrichment in the later timepoints, indicating a strong competitive advantage of Myc-activation, specifically in the pancreas. To confirm the pro-tumorigenic role of Myc in the pancreas, we analyzed the survival and phenotype of Myc-activated PPKS tumors, which revealed accelerated tumor progression relative to controls. Conclusions: We present one of the first implementations of autochthonous in vivo CRISPR activation screening, allowing for direct observation and quantification of oncogene selection and competition in vivo. Our results replicate tissue-specific differences in oncogene selection and indicate a dominant role of Myc-activation in pancreatic oncogenesis. Our platform has the potential to be used to address a wide range of inquiries into tumor biology, including drivers of progression, metastasis and resistance to targeted therapies.