Sara van den Berg

209 8 General discussion to a heterologous infection. Whether the claim of CMV-enhanced ageing on the influenza antibody response represents outliers based on chance, or instead a true, small, negative effect of CMV in certain populations or individuals remains to be investigated. LARGE EFFECTS OF CMV ON THE T-CELL POOL: DETRIMENTAL CHANGES, OR THE REQUIRED CMV IMMUNE RESPONSE? CMV-induced changes in the T-cell pool are no direct proof of CMV- enhanced ageing of the immune system The many parallels between age-induced and CMV-induced changes in the T-cell pool have laid the basis for the CMV-enhanced ageing theory. In CMV-enhanced ageing research, hallmarks of an aged T-cell pool, for example increased expression of senescence markers, are sometimes used as a read-out of a CMV-impaired immune system. However, CMV induces large numbers of CMV-specific T-cells that express high levels of senescence markers [42]. Consequently, there is an increased expression of senescence markers, which resemble the changes that occur with age in the T-cell pool. Investigating the effect of CMV on the T-cell pool changes that resemble the hallmarks of CMV-specific T-cells is like investigating the effect on the alcohol level when adding vodka to wine. The fact that CMV-specific memory T-cells are present in high frequencies (‘lots of vodka’) and themselves express high levels of senescence markers (‘high alcohol’) is often not taken into account when interpreting the overall level of immunosenescence (‘the alcohol level in the mixed drink’). Since the specificity of the T-cells is not taken into account, the increased level of senescence in the T-cell pool is taken as a ‘bad thing’, regardless of the specificity of the cells. However, possible associations between the frequency of for example senescent T-cells and an immunological outcome should really be based on senescent T-cells specific for the immunological challenge. It is therefore crucial to investigate not only general immunosenescence levels in the T-cell pool, but specifically investigate non-CMV-specific memory T-cells as well. Does CMV affect the phenotype of non-CMV specific memory cells? It has long remained unknown to what extent CMV-specific T-cells themselves are responsible for the CMV-induced changes in the T-cell pool. Studies have suggested a change by CMV infection in phenotype of other antigen-specific T-cells like EBV-specific T-cells [30]. Unfortunately, influenza-virus specific T-cell numbers ( chapter 4 ) were too low to reliably measure their expression of senescence markers. However, other work from our group on EBV-specific T-cells, showed no changes in the expression of senescence markers by CMV infection ( Lanfermeijer et al , manuscript in preparation). Our t-SNE analysis did suggest that the CMV-induced alterations in the T-cell pool are not fully explained by the presence of CMV-specific T-cells ( chapter 6 ). The increased T-cell numbers with high senescence marker expression in CMV+ individuals were only to some extent overlapping with the phenotype of CMV-specific T-cells. Note that a t-SNE analysis is based on the expression level of markers (continuous outcome: mean fluorescence intensity), while most papers merely investigate presence of or absence of a marker (dichotomic outcome: negative or positive) [30] (Lanfermeijer et al, manuscript in preparation). So differences in non-CMV