Evolution and Immunity in Cancer and HIV

Abstract

Cancer and HIV are frequently-incurable diseases with high global burden. As evolving populations within a single individual, both exhibit dramatic expansions in size and diversity which interact with the immune system and complicate treatment. The advent of cheap and accessible DNA sequencing and quantification technologies has enabled detailed measurements of these diseases from samples often distributed sparsely in time. Integrating these data into a quantitative, dynamical view of disease progression in order to improve treatments remains an open challenge. In this thesis, we explore examples of data integration with dynamical models in three areas: cancer evolution, cancer surveillance by the immune system, and HIV infection under an immune response. In the first chapter, we describe a tool for phylogenetic inference using DNA sequencing of spatially distinct samples from a cancer. In benchmarks, the tool overcomes noise introduced by sequencing to provide a picture of the evolutionary history of a tumor. In the second and third chapters, we describe the results of applying this tool to two datasets, first in primary pancreatic cancers with matched preneoplastic lesions and next in untreated metastases. We find many shared driver mutations among the primary tumor and preneoplastic lesions, suggesting preneoplastic cells can spread through the pancreas. In untreated metastases, we observe limited driver gene heterogeneity, consistent with a model of growth, mutation, and metastasis seeding from the primary tumor. In the fourth chapter, we describe a branching process model of neutral evolution in tumors and fit analytical predictions from it to cancer sequencing data. This neutral model explains patterns in sequencing data for many tumors. In the fifth chapter, we propose and analyze a simple model of cancer immune surveillance. We find that tumors susceptible to immune clearance must have a rate of mutation higher than is usually observed clinically. In the final chapter, we propose a dynamical model of viral rebound and immune control and compare it to data from several studies in macaques infected with SIV and SHIV and treated with immunotherapy. These results are combined with data from HIV infection to make predictions for future trials in humans.