Sara van den Berg

186 Chapter 7 The higher CD4+ T EM and T EMRA numbers in ST CMV+ individuals were not related to sex or age at end point (data not shown). Since ST CMV+ individuals were significantly older at CMV seroconversion than LT CMV+ individuals (average of 57 years and a maximum of 43 years respectively, p<0.0001), the association may be driven by differences in age at seroconversion. Indeed, individuals who seroconverted at older age (≥45 years old, n=18/19) had significantly higher CD4+ T EM and T EMRA cell numbers at end point than individuals who seroconverted when they were younger than 38 years of age (n=26/116 LT CMV+ individuals, p=0.018 and p=0.006 respectively, Figure 4C ). However, within the group of ST CMV+ individuals, age of seroconversion was not significantly related to CD4+ T EM or T EMRA cell numbers (data not shown), possibly due to the smaller age range. Within the group of ST CMV+ individuals, we found a positive correlation between duration of CMV infection and CD4+ T EM and T EMRA cell numbers at endpoint ( Figure 4B ). Similar results were observed for the differentiation states of T EM and T EMRA cells (data not shown). Together, these results suggest that duration of CMV infection has only a minor effect on CMV-induced CD4+ T EM and T EMRA cell numbers, and that becoming CMV+ at older age leads to higher CD4+ T EM and T EMRA numbers at end point. We further investigated the functionality of the CMV-specific CD4+ T-cells of ST CMV+ and LT CMV+ individuals. CD4+ T EMRA cell numbers at endpoint correlated positively with the percentage of granzyme B and perforin positive T EMRA T-cells ( Figure 4D ) and with the CMV-specific IFN γ response after stimulation with overlapping CMV-peptide pools (UL55, pp65 and IE-1) ( Figure 4E ), suggesting that CD4+ T EMRA cells have cytotoxic potential and are responding to CMV by cytokine production. After stimulation of CD4+ T-cells with overlapping CMV peptide pools, ST CMV+ individuals showed higher IFN γ -production than LT CMV+ participants ( Figure 4F ). CD4+ CMV-specific T-cell responses were polyfunctional, producing IFN γ , TNF α , MIP-1 β and CD107 but not IL-2, suggestive of an end-stage highly functional T-cell phenotype. Higher percentages of these cells were found in ST CMV+ compared to LT CMV+ individuals ( Figure 4G, Supplementary Figure 4C) . This was mainly due to the fact that ST CMV+ individuals had significantly higher polyfunctional responses to peptide pool UL55 than LT CMV+ individuals (data not shown). These results suggest that many of the CD4+ T EM and T EMRA cells in CMV-infected individuals are CMV-specific T-cells with a polyfunctional late-stage memory phenotype, and that these cells are induced early after primary CMV infection. In addition, these cells are induced in higher numbers in individuals who became CMV infected at older age. The median frailty index score of all participants was 0.069 at round 5 (Inter Quartile Range (IQR) 0.027– 0.167) and 0.081 (IQR 0.029 – 0.186) at round 6, indicating that on average frailty increased with age (p<0.001). Median increase in frailty index was 0.013 (IQR -0.017– 0.049) (p<0.001) ( Supplementary Figure 5 ). We found no significant differences in frailty between CMV-, ST CMV+, or LT CMV+ individuals ( Figure 5A ). While there seemed to be a negative trend between frailty and duration of CMV infection within ST CMV+ participants (n=19), this correlation was not statistically significant (p=0.098, ρ =-0.27 Figure 5B ).