Use of retroviral integration to study evolution and cancer

Abstract

Somatic infection by retroviruses can cause neoplastic disease if proviral formation results in alteration of the expression of a cellular cancer gene. Alternatively, if a retrovirus infects the host germline, it can be transmitted like any other Mendelian locus. Retroviruses in the germline are known as endogenous retroviruses.;My work has focused on two consequences of retroviral integration. The first was a screen to identify oncogenes that cooperate with a loss of function at the Cdkn2a locus to cause cancer. The Cdkn2a locus encodes two unrelated tumor suppressor proteins, p16 Ink4a and p19ARF. I infected Cdkn2a (-/-) mice with Moloney Murine Leukemia Virus (Mo-MuLV), which causes tumors by insertional activation of cellular oncogenes. PCR, DNA sequencing and analysis of the mouse genome were used to identify novel candidate oncogenes.;Cdkn2a (-/-) mice infected with Mo-MuLV developed lymphomas and myeloid tumors. 747 distinct viral integration sites were identified from a panel of 104 tumors. Analysis of these integration sites yielded 46 common integration sites (CIS), 37 of which have not previously been described.;The second aspect of my work was an assessment of the replicative potential of Human Endogenous Retrovirus (HERV) K. We found that most of the full-length HERV-K proviruses in the human genome formed after the human lineage separated from those leading to chimpanzees and gorillas and that most proviruses are fixed in the human genome. We identified two proviruses that integrated so recently that they are allelic with the preintegration site.;I was also interested in examining the coding capacity of the various viral ORFs as they exist in the genome today. We found full-length ORFs for each of the viral primary translation products to be present in multiple individual proviruses. We also identified two HERV-K proviruses that contain full-length ORFs for all viral proteins.;We identified a HERV-K provirus that is present in chimpanzees and gorillas but not in humans. Our result requires that the divergence of gorillas from the human-chimpanzee common ancestor occurred a sufficiently short time before the divergence of the latter from each other that allelism was maintained throughout the period of existence of the human-chimpanzee common ancestor.