| Summary: | Public health strategies to tackle HIV-1 epidemics in a variety of national settings need to be aware of the factors involved in transmission of infections. There is support in the literature for the hypothesis that individuals with recent infection may contribute disproportionately to onward transmission due to the high viral load and lack of infection status associated with this stage. This thesis sets out to further explore this risk group, and to develop methods to monitor its impact on epidemics. Two epidemic settings, the United Kingdom and Kumasi, Ghana, were investigated using molecular epidemiological techniques to assess the role of individuals with recent infection in the formation of transmission clusters. A classifier of HIV-1 infection length was developed based upon the proportion of mixed nucleotides within consensus pol gene sequences and applied to phylogenies constructed using viral sequences obtained from each cohort. In the Ghanaian setting, the performance of the nucleotide ambiguity classifier was compared to an antibody avidity based measure of infection length, to gauge the usefulness of both approaches in a sub-Saharan setting. In order to more fully explore the complexity of intra-host HIV-1 quasi-species dynamics in the early phase of infection, a well defined cohort of UK-based individuals, some with multiple pre-treatment time points, had their virus deep-sequenced and analysed using a subpopulation reconstruction approach. The proportion of recent infections identified within the UK HIV-1 epidemic by the classifier agreed well with previous studies. Application of the classifier to UK-wide phylogenies revealed disproportionate linkage of recently infected individuals to clusters across subtypes. Phylogenetic analysis of the Kumasi cohort did not reveal a highly clustered epidemic. Transmitted drug resistance was present at a level consistent with reports elsewhere in West Africa. Comparison of the antibody avidity based measure of infection length with the nucleotide ambiguity classifier indicated both markers co-segregate together, but produced differing estimates of the proportion of recent infections in the cohort. Deep-sequencing of recently infected individuals revealed some individuals may have been infected with more than one viral subpopulation, whilst others appeared to have been infected with a single subpopulation. This work supports the utility of a consensus sequence based measure of infection length in assessing the role of recently infected individuals in driving epidemics on a large scale. Such a measure will need further refinement and validation depending on the setting used, but presents a potentially useful biomarker that could be used in conjunction with other clinical parameters. Deep-sequencing of HIV-1 in recently infected individuals points to the quasi-species complexity that exists between and within individuals, and the subpopulation reconstruction approach taken in this work reveals dynamics at play which have the potential to impact on vaccine design and molecular epidemiological monitoring of epidemics.
|
|---|
