Human cytomegalovirus (HCMV) is a betaherpesvirus that establishes a lifelong infection in hosts. In the majority of cases, the immune response to primary HCMV infection limits viral replication and dissemination, such that overt clinical disease is prevented. However, the immune system cannot prevent the virus establishing a latent infection, which enables lifelong persistence. Historically, a long-standing view was that viral gene expression during latency was largely absent, thus facilitating the avoidance of immune detection. However, it has now been established that viral activity in latency is far from quiescent, and the expression of a number of viral genes is known to occur. Therefore, an important question arises: are T cells specific to these proteins generated, and, if so, why are HCMV latently infected cells maintained in the face of these potential T cell responses? Previous work has shown that two such viral gene products, UL138 and LUNA, are recognised by CD4+ T cells, with a subpopulation of these cells secreting the immunosuppressive cytokines interleukin (IL)-10 and transforming growth factor beta (β). However, little is known about the host immune response to other key latency-associated viral proteins; US28, UL111A, and UL144. Using overlapping peptide pools designed to cover the whole of the predicted amino acid sequence of these HCMV proteins, in combination with fluorescent ELIspot (FluoroSpot), ELISA, and intracellular cytokine staining, I have determined the frequency, cytokine secretion profile, effector function, and memory phenotype of CD4+ and CD8+ T cells in a large cohort of HCMV seropositive healthy donors. My results show that these viral gene products are also recognised by CD4+ T cells and are composed of distinct cellular populations secreting either IFNγ or IL-10. The high sensitivity of this assay has also revealed previously uncharacterised CD8+ T cell responses to US28, UL111A, and UL144, as well as responses to LUNA and UL138. Intriguingly, IL-10 secretion by a distinct population of latency-specific CD8+ T cells was also observed. T cell responses to latency-associated ORF products were found to be composed of greater proportions of IL-10 secreting cells compared to responses to the lytic ORFs pp65, IE1, and gB. The frequencies of IL-10 secreting T cells specific to HCMV latency-associated proteins did not increase with greater time of viral carriage, as measured indirectly by donor age. Although IFNγ secreting T cells specific to latency-associated proteins were detected in kidney transplant recipients immediately following primary HCMV infection, there were no such IL-10 secreting sub-populations, despite IL-10 T cell responses being detected to two lytic proteins, US3 and pp71. Given that latency-specific IL-10 secreting T cells were found to be a separate population to those secreting IFNγ, it was hypothesised that depleting the IL-10 secreting cells could improve the recognition and killing of latently infected cells by the specific but non-IL-10 producing T cell population. Classic regulatory T cells (Tregs) can be defined as CD4+CD127^lo CD25^hi, and depleting CD4+CD25^hi cells was investigated as a means to remove the latent-specific IL-10 secreting T cells. The results showed that this had a variable effect, suggesting that HCMV-specific IL-10 secreting cells of this Treg phenotype represent only a proportion of the IL-10 secreting CD4+ T cell population. As this strategy could not provide a consistent method to remove IL-10 secreting CD4+ T cells, neutralising the effect of IL-10 and TGFβ using anti-cytokine and anti-receptor antibodies was tested. My initial data suggest that treatment of latently infected cells with these neutralising antibodies in the presence of CD4+ and CD8+ T cells reduced latent viral carriage.