Limits of viral adaptation to the antigen presentation pathway

The vertebrate immune system uses several strategies to combat viral infections. Cellular immune responses evoked by the cytotoxic T cells (CTLs) are initiated after screening host cells for abnormalities. Examples are peptides derived from virus (epitopes) that are presented on the cell surface by major histocompability molecules (MHCs). Viruses therefore face strong selection pressures to evade the pathways signaling such abnormalities to the immune cells. In this thesis we address the co-evolutionof pathogens and hosts of one such a pathway, i.e., the classical Ag presentation pathway. The recent adaptation of HIV-1 to its new human host was studied by enumerating the total number of predicted CTL epitopes and epitope precursors in a HIV-1 sequences gathered over more than 25 years. We investigated the adaptation of HIV-1 by comparing the distribution of CTL epitopes in HIV-1 proteins with other proteins, and found no conclusive evidence for any ongoing accumulation of CTL epitope escape mutations in HIV-1. This was surprising because two important molecules in the pathway (proteasome and TAP) display hardly any genetic variation between individuals, and a rapidly evolving virus like HIV-1 should in 25 years be able to evolve an escape from such monomorphic molecules. We hypothesized that the virus fails to escape from the monomorphic steps of the pathway because the most specific molecule in this pathway, i.e., the MHC, is highly polymorphic (i.e., is genetically diverse in the human population). Upon transmission of virus to a new host, previous proteasome and TAP escape mutations are released from immune selection pressure because the new host's MHC selects another set peptides from HIV-1's proteome than the previous host. This scenario was tested in an individual-based computer simulation model, in which an HIV-1 like virus was allowed to adapt to a human like host population. We showed that different degrees of MHC polymorphism in the host population affected how much the virus was adapting to the polymorphic MHC, or to the monomorphic proteasome and TAP. Finally, we extended the model such that the host population could evolve all components of its antigen presentation pathway in response to multiple endemic pathogens. This allowed us to study whether the current structure of the human antigen presentation pathway could be explained just in terms of host-pathogen co-evolution. Based on our earlier hypothesized mechanism, we expected that only the most specific step of the pathway would evolve a polymorphism. Under the constraint that the individual steps of the pathway remain co-evolved, it was indeed only the most specific step of the Ag presentation pathway that became polymorphic