Identification of genotype-specific dependencies in Keap1-deficient lung adenocarcinoma

This electronic version was submitted by the student author. The certified thesis is available in the Institute Archives and Special Collections.Thesis: Ph. D., Massachusetts Institute of Technology, Department of Biology, 2019Cataloged student-submitted from PDF version of thesis. Vita.Includes bibliographical references.Lung adenocarcinoma (LUAD) remains the world's leading cause of cancer related mortality with an estimated two-hundred thousand new cases arising in 2019 in the United states alone. Molecular characterization of patient tumors has identified activating mutations in the small GTPase, KRAS, encompassing ~30% of all LUAD patient samples. Efforts in therapeutic intervention for KRAS-mutant LUAD are numerous, but have been met with little clinical success. Importantly, large-scale sequencing studies have defined the genomic complexities of LUAD resulting in the need to functionally characterize frequently mutated genes in the context of tumorigenesis. The development of somatic genome editing provided by clustered regularly interspaced short palindromic repeats (CRISPR) and the endonuclease, CRISPR associated protein 9 (Cas9) has accelerated our ability to functionally interrogate putative genetic drivers in the context of mammalian cells and autochthonously arising tumors.Using CRISPR/Cas9, we validate the redox sensor, Keap1, as a functional tumor suppressor that is frequently mutated in human LUAD. Notably, loss of Keap1 results in the hyperactivation of the Nrf2-regulated antioxidant response program with a concomitant increase in lung tumor burden and advanced disease. Moreover, we find that Keap1-mutant tumors display a heightened dependency for glutaminolysis. Pharmacological glutaminase inhibition suppresses the growth of Keap1-mutant tumors and human KEAP1-mutant patient-derived xenografts (PDXs). In an expanded CRISPR/Cas9 genetic screen targeting the druggable genome, we identify a novel Keap1-mutant specific vulnerability to loss of the putative endoplasmic reticulum resident acetyl-CoA transporter, Slc33a1. Targeted in vivo somatic ablation of Slc33a1 results in decreased tumor burden in Kras-driven; Keap1-deficient genetically engineered mouse models of LUAD.In short, the data presented here demonstrate the power of CRISPR/Cas9 somatic editing to functionally validate tumor promoting and tumor suppressive genetic events with the simultaneous identification of genotype-specific vulnerabilities for a rapid target identification pipeline.by Rodrigo Romero.Ph. D.Ph.D. Massachusetts Institute of Technology, Department of Biolog