Sara van den Berg

144 Chapter 6 likelihood approach. Using a bootstrap method, the 95% confidence limits of the parameter were determined by resampling 500 times the residuals to the optimal fit. RESULTS Characterization of the study cohort We included 54 healthy older adults of on average 68.8 years of age (SD 4.6 years, range 61.0 to 79.8 years) who were free of systemic diseases or immunocompromising conditions. Individuals with any (history of) immune-mediated diseases or any current medicine use (with the exception of occasional use of paracetamol or ibuprofen) were excluded. Using a multiplex immunoassay to test for CMV antibodies, we identified 32 of the 54 individuals (59%) as CMV+ and 22 as CMV- ( Supplementary Table 1 ). No significant differences in sex or age were observed between CMV+ and CMV- individuals (data not shown). We investigated the influence of CMV-seropositivity on the T-cell pool by measuring absolute T-cell numbers, and their memory or senescence-associated phenotype in CMV- and CMV+ individuals. In line with previous reports [2, 3, 30, 31], we found significantly increased cell numbers of CD4 + and CD8 + T-cells with a T EM or T EMRA phenotype in CMV+ compared to CMV- individuals ( Figure 1A ). Moreover, CMV-seropositivity was associated with a significantly increased expression of the senescence markers CD57 and KLRG-1 and a loss of expression of the co-stimulatory receptor CD28 on CD4 + T EM , CD4 + T EMRA and CD8 + T EM cells, identifying a late-stage differentiation state ( Figure 1B ). For CD8 + T EMRA cells, the differences in CD57, KLRG-1, and CD28 did not reach statistical significance, but similar trends were observed. In addition, 38 CMV-specific CD8 + T-cell populations were characterized in 32 CMV+ participants using HLA-class I tetramers for immunodominant CMV epitopes, based on the HLA-type of each participant (see Supplementary Table 1 ). The frequencies of CMV-specific CD8 + T-cells specific for one epitope reached on average 3.7% of total CD8 + T-cells and could reach up to 15-20% for some epitopes ( Figure 1C upper panel ). The dominant phenotype of CMV-specific CD8 + T-cells varied between individuals, although mosT-cells had a T EM / T EMRA phenotype ( Figure 1C middle panel ). Interestingly, we observed a significant positive correlation between the frequency of CMV-specific CD8 + T-cells and the percentage of CMV- specific CD8 + T-cells with a T EM/EMRA phenotype ( p = 0.035, r = 0.37) ( Figure 1C lower panel ). Furthermore, CMV-specific CD8 + T-cells showed a significantly higher expression of the markers CD57 and KLRG-1 and a lower expression of the receptor CD28 than bulk memory T-cells of CMV+ individuals ( Supplementary Figure 2A ). Even within the different memory subsets, CMV-specific CD8 + T-cells showed a more pronounced late-stage differentiation state compared to the total memory subset, especially within the T CM and T EM compartment ( Figure 1D ). Thus, CMV-specific CD8 + T-cells were the most differentiated, followed by bulk memory CD8 + T-cells in CMV+ individuals and then bulk memory CD8 + T-cells in CMV- individuals.