Targeting Non-Genetic Mechanisms of Therapy Resistance in Cancer

Abstract

The malignant transformation of normal cells to cancer is driven by genetic and epigenetic mechanisms that allow for cellular immortalization, uncontrolled proliferation, immune escape, and increased migratory potential. Although the therapeutic agents we employ to eliminate cancer cells capitalize on inherent differences in malignant vs. normal cells, drug resistance is still a common reason why cancer patients succumb to the disease. Cancer populations that acquire resistant properties emerge from residual cancer cells that endure through the initial therapy via genetic and non-genetic survival mechanisms. Much of our academic and industry resources have focused on identifying and drugging genetic mutations in resistant tumors through large-scale DNA sequencing studies in resistant tumors. Although specific gatekeeper mutations in resistant tumors have been identified, we are beginning to understand that the complexities of drug resistance go well beyond purely genetic alterations. In fact, one of the most convergent properties of drug resistant tumors are transcriptional cell state changes governed by epigenetic modulation. Despite this fact, very few therapeutic targets of drug-resistant transcriptional cell states have been uncovered. The work in this dissertation broadly focuses on identifying therapeutic opportunities created by the non-genetic transcriptional states that drive drug resistance with the following three projects. First, we review our recent discovery that cancer cells surviving treatment with cytotoxic cancer therapies sustain a sublethal level of caspase activation that contributes to the phenotypic properties of drug-tolerant persister populations. Using CRISPR/Cas9 knockout cancer models, we found that caspase activation following mitochondrial outer membrane permeability (MOMP) blocks inflammation in cancer cells following MOMP. Specifically, in the presence of caspase-inhibition, we found mitochondrial RNA (mtRNA) leaks into the cytosol post-MOMP and activates the mitochondrial antiviral-signaling protein (MAVS) to stimulate Type I Interferon (IFN) production. With syngeneic in vivo models, we discovered that blocking caspase activity in chemotherapy-treated tumors stimulated an anti-tumor CD8+ t cell immune response and improved survival outcomes. These data suggest that inhibiting caspases in DTPs may reawaken the immune system against these residual cancer cells through mtRNA-MAVS-IFN signaling. Next, we investigated how lineage differentiation states contributed to the emergence of drug dependence in BRAFV600E melanoma models. Interestingly, through unbiased transcriptomic analysis, we found melanocytic differentiated BRAFV600E melanoma was the most likely subtype to become dependent upon continuous BRAF-inhibitor (BRAFi) therapy for survival in resistant models. Specifically, we found that loss of the lineage transcription factor Microphthalmia-associated transcription factor (MITF) drove drug dependence in melanocytic melanoma models. We exploited this discovery using simple mathematical models to identify an optimal intermittent therapy schedule to minimize population growth. Finally, we systematically identified the dependency network of EMT using gene essentiality scores in 873 cancer cell lines and nominated a rho-effector kinase, PKN2, as a top therapeutic target of the mesenchymal-like state. Mechanistically, we found that PKN2 promotes the survival of mesenchymal-like cancer through its positive regulation on the oncogenic transcriptional activator WW domain containing transcription regulator 1 (TAZ). Using state-of-the-art quantitative phosphoproteomics, we discovered that PKN2 directly antagonizes the hippo tumor suppressor pathway through phosphorylation of SAV1S90, which resulted in downstream TAZ activation. Additionally, using the National Cancer Institute's Clinical Proteomic Tumor Analysis Consortium (CPTAC) dataset, we found that PKN2 activity was enriched in patients with mesenchymal-like features. Translationally, we inhibited PKN2 in mesenchymal-like residual cancer cells and found it eradicated the persister cells and prevented drug resistance.