Setting aside the difficulties in proving that accumulation of cytomegalovirus-specific T cells with the distinct phenotypic and functional characteristics described above are responsible for an increase in all-cause mortality or a decreased ability to mount good quality responses to new antigens, and whether their specificity is important in this context, both Hill (Portland) and Klenerman (Oxford) addressed the biological phenomenon of "memory T cell inflation" using the murine CMV model. Klenerman has investigated the mechanism that leads to memory inflation while Hill has asked what happens to the CMV-specific T cell response if cytomegalovirus replication is inhibited. This latter investigation is of particular importance as it has been suggested that elderly HCMV-seropositive people may benefit from anti-herpes virus chemotherapy. The rationale would be that anti-viral therapy would decrease HCMV replication and thus chronic antigenic stimulation which might reduce the frequency of the late-stage differentiated T cell pool and thus allow an increase in the naïve pool with the hope that T cell responses to neo-antigens would improve. Klenerman addressed the issue of why memory inflation is triggered by some antigens but not others. Memory inflation is defined as the inordinate increase in frequency of CD8+ T cells specific for certain antigens with time post-exposure to antigenic stimulation. The phenomenon has been seen with CMV, HIV and parvovirus B19 infections; however, certain antigens from these viruses but not to others within the same virus are found to cause this effect, and other viruses do not cause it at all. The "inflating" cells have a distinct transcriptional profile as compared to non-inflating T cells from the same animal. It is not clear if persistence of viral antigen is required regardless of viral reactivation or why some epitopes cause inflation while others do not. The Klenerman study concluded that viral replication was not required for memory inflation and that epitopes that showed inflation had a reduced dependence on the immunoproteasome . Hill also addressed memory inflation in MCMV specifically to understand what happens to the T cell response if cytomegalovirus replication was inhibited? Using two separate models of replication and spread-deficient virus that still establishes viral latency, memory T cell inflation nonetheless occurred and was maintained. Inhibition of viral replication by inhibiting viral DNA replication also had no effect on the T cell response over time .
Taking these two studies together, the data from the murine model would suggest that treating humans with anti-herpesvirus chemotherapy may not reverse the effects of HCMV on the memory T cell compartment.
Most studies of this type have been carried out in humans or mice, but Nikolich-Zugich (Tucson) presented data looking at the maintenance of T cell memory in the face of CMV and ageing using the Rhesus CMV (RhCMV) model. The study examined the RhCMV CD4+ and CD8+ T cell responses in adult animals aged 7-10 years and aged animals 19-26 years old. The study concluded that there was no difference in the phenotype, numbers, proliferative capacity or functionality of the T cells between the two groups of animals and that there was no evidence for either a functional decline or exhaustion in aged animals . This was further corroborated by their study using systemic HSV-1 infection, which closely models MCMV infection in mice . However, given that CMV and HSV have very different effects in humans , this result in monkeys is perhaps not unexpected.
Further, in the context of the impact of HCMV on immune signatures, Sauce (Paris) presented her studies on the impact of HCMV infection in young adults thymectomized during early childhood. The data presented suggested that HCMV infection tended to cause the exhaustion of the naïve T cell pool . Moreover, in a separate study in HAART-treated HIV patients, mounting strong CMV-specific responses appeared to impact on both naïve CD4+ T-cell counts and their recovery upon antiretroviral therapy . In these two clinical settings, T cell responses to HCMV have a profound impact of the immune system as a whole. In the special cases of thymectomy and HIV infection, the impact of CMV may therefore be more pronounced and easier to determine unequivocally.
Further discussion considering the impact of CMV on immunosenescence led to revisiting the question of what is the definition of T cell senescence? Senescence is a term generally used to describe replication incompetence of fibroblast cells after a certain finite number of population doublings. Replicative senescent cells have been defined in the brain and in cardiology but how should it be done for the immune system? Are there any such cells in the constantly-renewing T cell compartment? The use of the term in describing CMV-specific T cell populations is more controversial as it is acknowledged that many of the late-differentiated T cells do retain replication competence. However, it should be noted that their definition by CD45RA expression and lack of either CD27 or CD28 or both does not define a homogeneous cell population, as many more markers can be added (eg CD57, KLRG1, PD1) to further sub-divide late-stage T cell populations which may yet include an antigen-specific population that is indeed unable to divide further. Another complication is that the term "senescent" is also being used to imply non-functionality, whereas data from many laboratories in human, rhesus and murine systems have shown that late-differentiated cells maintain function. In vivo, the work of Macallan (London) suggests that these cells turn over only very slowly. There was a feeling that we might have become too fixated on the replication capacity of these cells. Of more importance could be their direct effector function. Indeed, the suggestion was made (Pawelec, Tübingen) that some of the defining characteristics of senescent cells which T cells share with the "classic" fibroblast model (ie. lack of proliferative capacity, apoptosis resistance, inflammatory mediator secretion) are not markers of senescence in the sense that they are detrimental (which by definition of the term "senescence" has to be shown to justify its use). Instead, one could view these properties as cellular protection mechanisms responsible for maintaining an active population of cells required for immunosurveillance, ie. required for prevention of peripheral clonal deletion of necessary cells. In fact, there is some evidence from the longitudinal NONA studies that at the terminal phase of life in the very elderly, the number of CD8+ clonal T cell expansions, most of which are specific for CMV, is inversely related to remaining survival time . More recently, data from a 7-year follow-up of subjects 88 years old at baseline in the Leiden 85+ study has shown that the degree of accumulation of late-differentiated CD8 cells, putatively mostly specific for CMV, is positively associated with longer survival (Derhovanessian et al., 2011 submitted).
Discussion of CMV and Immunosenescence - Is it associated at all?
If chronic challenge with CMV contributes to causing immunosenescence, it might be expected that more frequent CMV reactivation would accelerate this process. However, it is not known whether CMV does reactivate more frequently in the elderly. It was felt that higher HCMV IgG levels might reflect this, but in many sick people they stay the same, so did this mean that HCMV was not reactivating? A study of elderly people suffering from acute bacterial infections (e.g. S. pneumonia, E. coli, S. aureus, K. pneumonia) did not show any reactivation of CMV infections measured by IgM antibodies (n = 20). However, serology does not measure viral levels so this is a mute point. Nonetheless, it was strongly felt by a number of clinicians involved with the care of older adults that "It is very unlikely that CMV reactivation and viral replication is a problem in the very elderly." Wang (Baltimore) provided some evidence for this, indicating that a small-scale study of older adults (n = 40) categorised as either frail or not frail showed no HCMV DNA in serum in either group, with the exception of one frail individual who had 8 copies/mL (this would be on the very edge of detection). People were reminded that VZV reactivation occurs in the elderly and that it had been correlated with decrease in CD4 cells and decrease in IFN gamma responses . Data on CMV are lacking.