Sara van den Berg

155 6 T-cell dynamics in CMV that are generated by CMV infection is unparalleled by any other antigenic stimulus, and CMV-specific T-cells display a unique late differentiated phenotype [35]. In this study, we investigated the effect of CMV infection on the dynamics of the T-cell pool in healthy older adults. We show that the CMV-induced late-stage differentiated state of the CD8 + T-cell memory pool is not explained by the sheer presence of large numbers of CMV-specific CD8 + T-cells. In the CD4 + memory T-cell pool, latent CMV infection is associated with increased expression of late-stage differentiation markers and decreased proliferation, supported by a marked decrease in Ki-67 expression and a trend towards lower T-cell production rates assessed by in vivo deuterium labelling. Despite clear differences in the expression of late-stage differentiation markers, we find no significant differences in the expression of proliferation and apoptosis markers or in the in vivo production rates of different CD8 + memory T-cell subsets between CMV- and CMV+ individuals. CMV-specific CD8 + T-cells also express a late-stage differentiated phenotype, but do not differ significantly in in vivo production or loss rates when compared to T EM/EMRA or T CM cells in CMV- or CMV+ individuals. Finally, we find that the expression of late-stage differentiation markers correlated negatively with T-cell production rates, coupling high expression of senescence markers to an increased lifespan in vivo . Originally, the typical features of the T-cell pool in CMV+ individuals were thought to reflect an effect of CMV on the whole T-cell pool [30]. With the invention of MHC class I tetramers, however, it became clear that CMV-specific CD8 + T-cells are present in large numbers, and themselves possess a phenotype that resembles the changes observed in the total T-cell pool of CMV+ individuals. We therefore investigated whether the phenotypical changes of the T-cell pool in CMV+ individuals are the direct consequence of the presence of large expansions of CMV-specific late-differentiated CD8 + T-cells. By performing a t-SNE analysis, we found that CMV-specific CD8 + T-cells only partially explain the differences in the CD8 + T-cell pool between CMV- and CMV+ individuals. CMV infection thus seems to affect non- CMV-specific memory T-cells as well. Two different fluorochromes (PE and APC) for different epitopes of CMV-specific CD8 + T-cells yielded comparable results in the t-SNE analysis, indicating that the mean fluorescence intensity (MFI)-based clustering of the CD8 + T-cells was not biased by compensation for spectral overlap. The investigated epitopes in our analysis captured at most 18% of CMV-specific CD8 + T-cells out of the total CD8 + T-cell pool, while some studies suggest that in CMV+ individuals as much as 30-90% of CD8 + T-cells is CMV-specific [11, 12]. We cannot exclude the possibility that CD8 + T-cells specific for other (non-tested) CMV-epitopes may have been responsible for the observed differences between CMV+ and CMV- individuals. We regard this unlikely, however, because the 7 epitopes that were tested are the most immunodominant ones known for CMV. Ideally, a method that would pick up all CMV-specific CD8 + T-cells would need to be used to reveal the precise contribution of CMV-specific T-cells to the changes in the T-cell pool observed in CMV+ individuals.