| Summary: | Checkpoint blockade immunotherapies have transformed the standard of care and outcomes for many cancer types; however. more than 60% of patients still do not experience a durable clinical response from these treatments. To address this problem, the development of novel biomarkers and more effective combinatorial therapies are needed. In this thesis, we first explore and validate the use of extracellular vesicular (EV) RNA as a potential biomarker for immunotherapy response. We discover differentially expressed genes and pathways within the plasma-derived EV RNA that is concordant with known biology. We also show that mutational information contained within EV RNA can stratify responders and non-responders. We leverage a Bayesian probabilistic model to deconvolve the tissue-of-origin of EV RNA transcripts, allowing greater interpretability for differentially expressed genes and pathways. Next, we performed large-scale epigenomics profiling in two cohorts of immunotherapy patients, and we discovered a non-responder enhancer signature that is lost in responders. Many genes contained within this epigenetic signature are associated with immunotherapy resistance, and we reasoned targeting this signature with acetylation-reader bromodomain inhibitors would allow suppression of multiple resistance mechanisms at once. We show that bromodomain inhibitors exhibit considerable synergism with anti-PD1 in reducing tumor volume in murine melanoma transplantation models, and this synergism also improves anti-tumor killing by tumor infiltrating lymphocytes. Using the same cohort, we also identify 189 peaks with differential activity in both the responders and non-responders, and we show these peaks are potentially predictive biomarkers of immunotherapy response. Finally, we leverage three transgenic mice lines to investigate the effect of T-cell receptor repertoire on cell fate commitment by CD4+ SP T-cells into either the thymic conventional (Tconv) or thymic T regulator (tTreg) lineages. We show based on overlap and machine learning analysis that T-cell receptors are not the sole determining factor in Tconv vs. tTreg cell fate decisions. Together, these projects offer new biomarkers and novel combinatorial treatment options for checkpoint blockade immunotherapies.
|
|---|