Date of Award
Doctor of Philosophy (PhD)
The majority of EGFR mutant lung adenocarcinomas respond well to EGFR tyrosine kinase inhibitors (TKIs). However, the majority of these responses are partial with drug tolerant residual disease remaining even at the time of maximal response. This residual disease serves as a reservoir for the emergence of acquired resistance and tumor relapse, which inevitably occurs in patients treated with TKIs. It is thus critical to understand the biology of residual tumor cells and find out the mechanisms that underlie drug tolerance. Knowledge of such mechanisms could lead to the identification of potential strategies to forestall the emergence of drug resistance. Studies of residual disease have been hampered by the difficulty of studying these persister cells in patient specimens and most studies to date have relied on analysis of established cell lines in culture. To investigate the cellular and molecular properties of residual tumor cells in vivo, I leveraged patient-derived models of EGFR mutant lung cancer. Three EGFR mutant PDXs were treated with the third-generation TKI osimertinib. Tumors regressed in all cases but measurable residual tumor remained in 2 out of the 3 PDXs even after 6 weeks of osimertinib treatment. Whole-exome sequencing (WES) of the untreated PDXs compared to the residual tumors revealed an unchanged mutational landscape between the samples indicating that genetic mechanisms did not account for drug tolerance. Bulk RNA-sequencing, however, revealed extensive transcriptional changes between the untreated PDX and residual disease. In one of the PDXs, I identified upregulation of the neuroendocrine lineage transcription factor ASCL1 in residual disease compared to untreated tumors. Using single-cell RNA-sequencing I found a pre-existing ASCL1hi tumor cell population in untreated tumors suggesting that these cells which possess drug tolerant properties were selected for during drug treatment. Depending on the cell line examined, expression of ASCL1 in human mutant EGFR lung cancer cell lines gave rise to persister clones following osimertinib treatment. This result demonstrated functionally that ASCL1 could lead to TKI tolerance and that whether it did this depended on the cellular context. Further gene expression profiling of ASCL1-transfected cell lines identified an ASCL1-induced epidermal-to-mesenchymal transition (EMT) signature that led to tolerance to osimertinib. Our studies provide insights into the role of the neuroendocrine factor ASCL1 as a potential driver of drug tolerance in mutant EGFR lung cancer and ongoing work is focused on identifying the mechanisms underlying the cellular context specificity of ASCL1-mediated EMT. In addition to PDX models, I also performed complementary studies on TKI tolerance in a genetically-engineered mouse model (GEMM) of mutant EGFR-driven lung adenocarcinoma. To identify and isolate drug tolerant persisters, we developed a transgenic mouse model in which mutant EGFR-expressing lung epithelial cells were labelled with a fluorescent marker, mKate. Using this model, I found that tumor-bearing mice responded almost completely to osimertinib and the number of mKate+ cells was decreased in response to TKI treatment and plateaued after 4 weeks of treatment. Targeted deep sequencing of the mutant EGFR transgene did not show enrichment of any resistance conferring mutations in mKate+ cells following TKI treatment, indicating that on-target EGFR mutations did not contribute to TKI tolerance in the GEMM. In contrast, RNA sequencing of DTPs sorted from this model revealed deregulation of metabolic and developmental pathways that could play a role in drug tolerance. In summary we have established and characterized state-of-the-art in vivo models to study drug tolerance and unveiled new potential mechanisms of TKI tolerance which serve as the foundation for future studies into this critical problem in cancer therapeutics.
Hu, Bomiao, "Identifying Mechanisms of Drug Tolerance in EGFR Mutant Lung Cancer" (2021). Yale Graduate School of Arts and Sciences Dissertations. 62.