Human Cytomegalovirus (HCMV) is a ubiquitous human herpesvirus. Primary infection with HCMV in healthy individuals is generally asymptomatic but the virus is never cleared and can go on to establish lifelong latency in cells of the early myeloid lineage. During latency, the virus reduces its gene expression programme and produces no infectious progeny virus, but maintains the ability to reactivate and cause a full lytic infection. In vivo, sporadic reactivation from latency routinely occurs but this is normally subclinical in healthy individuals. In contrast, HCMV primary infection or reactivation can cause severe disease in the immunosuppressed or immunonaïve, including transplant recipients and people living with AIDS. Although there are several drugs approved for the treatment of HCMV, none of them clear the virus from the host as they do not target latent infection. Consequently, reactivation from latency is a significant source of disease, and there remains an unmet need for treatments that also target latent infection. Therefore, understanding the processes involved in HCMV latency is of great importance.

In HCMV, RNA-Seq has identified two long non-coding RNAs that are amongst the most highly expressed viral genes during latency. Long non-coding RNAs are defined as RNAs exceeding 200 nucleotides in length that have little or no protein-coding potential. In this thesis, I have investigated the roles of these two long non-coding RNAs, RNA4.9 and β2.7, during latent infection. RNA4.9 has been implicated in epigenetic silencing of the viral genome during latency via recruitment of the repressor complex, PRC2, to the HCMV genome. I have tested whether knockdown of RNA4.9 impacts upon latent infection and present evidence that suggests it is not essential for the establishment of latency.

I have also investigated the role of β2.7 during latency and found that it protects against apoptosis induced by high levels of reactive oxygen species (ROS) in latently infected monocytes. In part, this is mediated by a substantial upregulation of antioxidant enzymes SOD2 and GPX-1. The absence of β2.7 can be rescued by the addition of an antioxidant. These observations demonstrate a novel way in which HCMV protects latently infected cells from pro-death signals to optimise latent carriage. Interestingly, I also demonstrate that the absence of β2.7 in infected monocytes leads to a lytic and not a latent infection. I show that the inability of a β2.7 deletion virus to silence viral genes and establish and maintain a latent infection in monocytes is related to control of ROS levels. Adding an antioxidant to β2.7 deletion virus-infected monocytes restores the silencing of viral genes seen in a latent wild-type (WT) virus infection of monocytes. Furthermore, I show that induction of ROS in WT-infected monocytes results in high levels of reactivation of latent virus. Taken together, this suggests that β2.7 plays an essential role in silencing the viral genome during latent infection, and that ROS have an important role in regulating viral gene expression during HCMV latency in monocytes.

Finally, given that reactivation of HCMV occurs under inflammatory conditions, which occur during infection with SARS-CoV-2, the causative agent of the ongoing COVID-19 pandemic, I tested whether lytic HCMV infection can affect the course of SARS-CoV-2 infection. Surprisingly, I found that HCMV can directly impact upon SARS-CoV-2 infection by upregulating the SARS-CoV-2 entry receptor, ACE2, resulting in increased SARS-CoV-2 infection in epithelial cells. These results warrant further clinical investigation as to whether HCMV infection influences the pathogenesis of SARS-CoV-2.