Single-Cell Genomics of Human Cancer

Abstract

Cancer is a complex disease involving not only malignant cancer cells but also adaptation and response by the immune system and stromal cells. Understanding the biology of the different types of cells comprising a tumor and how they influence each other is particularly necessary for understanding the evolution of resistance to therapy. Both cancer cell intrinsic and extrinsic resistance mechanisms are now understood to be necessary, and a solitary focus on only one component of a complexly evolving system is unlikely to yield comprehensive understanding. The advent and maturation of tools for single-cell genomic characterization of clinical specimens holds promise for understanding cancer therapeutic resistance. These methods allow isolation of biological phenomena that previously were obscured by bulk sequencing of samples with unknown cell type admixture. In doing so, they potentially enable dissection of how each of the component cell types of a human tumor behaves and is shaped by therapy on a scale that is unprecedented. While new technologies fuel new observations and hypotheses, they often also pose new challenges in interpretation. At present, single-cell genomics studies of clinical cohorts are often of limited scale, with highly sparse data and only rare longitudinal sampling. In cancer, this is further complicated by the different evolutionary histories of each patient’s disease, leading to a diversity of resistance mechanisms within patient cancer cells. As such, questions that may seem simplistic (e.g. What transcriptional changes are found in cancer cells after therapy?) remain major open questions. More complex questions of how the different component cell types may interact are even more challenging to address. In this dissertation, I use computational approaches in interpreting single-cell genomics and orthogonal data types to investigate response to therapy in advanced prostate and kidney cancers. The two disease types are treated with different therapeutic classes: the former with small molecules that target cancer cell-intrinsic biology and the latter with immune checkpoint blockade. In both settings, I show that multiple component parts of the tumor microenvironment change with therapy via dynamic interactions and provide hypotheses for further investigation and potential therapeutic development.