Synthetic lethal-based approaches to targeting KRAS-driven cancers have recently been receiving increasing attention. However, as KRAS activates multiple effector pathways, targeting single genes may not be sufficient to fully inhibit KRAS-driven oncogenesis. In addition, genome-wide screens can be limited by high noise and low signal. Furthermore, it is increasingly clear culturing cells in standard 2D tissue culture may not have limited capacity to evaluate important phenotypes relevant to cancer cells. We developed a multi-stage screening strategy that includes a) identification of Kras-relevant protein-protein interaction (PPI) networks, b) CRISPR-based screening of targeted pairs of genes that together are synthetic lethal and c)validation in a 3D sphere system using both human and mouse NSCLC cells.
We used affinity purification/mass spectrometry (AP/MS) to construct a detailed map of protein-protein interactions centered on KRAS. Based on this network we designed a CRISPR/Cas9 library targeting pairwise combinations of 119 KRAS-interacting genes. We screened two KRAS-driven Non-Small Cell Lung Cancer (NSCLC) cell lines (A549 and H23). This screen yielded a large number of gene pairs that synergistically impaired cell growth. We selected the 20 most promising targets for further screening in a panel of 5 KRAS-mutant and 4 KRAS wild type NSCLC cell lines. From these screens we identified a KRAS-specific synergistic interaction between two genes - Rap1GDS1 and RhoA.