Modelling the Evolution of HIV-1 Protein-Coding Sequences with Particular focus on the early stages of Infection

Abstract

Modelling the Evolution of HIV-1 Protein-Coding Sequences with Particular Focus on the Early Stages of Infection Natasha Thandi Wood, February 2010 The evolution of the viral genome sequence over the course of HIV-1 infection is of interest for vaccine and drug design, and for the development of effective treatment strategies. Characteristics of the transmitted viral genome that could render the virus more sensitive to host immune responses, are of particular interest for vaccine studies. However, sequence samples from the earliest phase of HIV infection are scarce, and inferences about the nature of the infecting virus and its evolution during the course of early infection are often made from samples isolated from later stages, or from chronic infections. To establish in detail the adaptive changes that occur in early infection, an investigation was carried out on a large dataset consisting of sequences isolated from individuals in early infection. The majority of these infections were inferred to have resulted from transmission of a single virion or virally infected cell, which permitted a detailed investigation of HIV-1 diversification in early infection for the first time. Comparing viral diversification across multiple patients, it was possible to identify specific evolutionary patterns in the HIV-1 genome that occur frequently during the earliest stages of infection. The analyses revealed that APOBEC-mediated hypermutation has an important role in early viral diversification and may enable rapid escape from the first wave of host immune responses. Several mutations in early infection that were likely to result in immune escape were identified, some of which have subsequently been confirmed experimentally. In general, experimental verification of model-based inferences is necessary, but can be expensive and time-consuming. To reduce the costs involved, it is essential that the evolutionary methods produce accurate results. Simulation results presented in this thesis show that inferences made about viral evolution can be subject to bias when key aspects of viral biology are not accounted for by the models used. In particular, some previous comparisons between sequence groups that share genealogical histories, positive selection studies that fail to account for recombination, and research on HIV covariation, may need to be revisited, using more accurate evolutionary models. The results presented in this thesis demonstrate the importance of accurate evolutionary models to understand the selection pressures acting on the virus during various stages of infection. Furthermore, using a phylogenetic model it was possible to identify sites in the HIV genome that were evolving adaptively and are implicated in CTL immune escape during early infection. Characterising escape mutations in the transmitted virus may lead to novel approaches to develop vaccines and antiviral drugs.