Effect of CMV on Immunosenescence
The cause-effect relationship between CMV and immunosenescence is not clear, despite the plethora of candidate contributory factors mentioned above. It is, however, clear that persistent CMV infection results in expansion of the total CD8+ T-cell pool associated with accumulations of late-stage differentiated effector memory T cells in the periphery. In CMV-seropositive people, especially older people, a remarkably high fraction of circulating CD8+ T lymphocytes is often found to be specific for CMV. However, although the proportion of naive CD8+ T cells is lower in the old than the young whether or not they are CMV-infected, the gross accumulation of late-stage differentiated CD8+ T cells only occurs in CMV-seropositive individuals . This phenomenon of gradual increase in CMV-specific CD8+ T-cell population over time is known as memory inflation [28, 29]. It is not clear whether this is adaptive or pathological; however, a lower fraction of CD8+ T cells bearing receptors for the HLA-A2- NLV epitope of CMV pp65 is able to produce IFN-y in response to specific peptide in the old than in the young. Nonetheless, the absolute number of T cells assessed as functional by this method is greater in the old than in the young, due to memory inflation . The large T-cell responses elicited by CMV are ascribed to numerous CMV proteins of which pp65 and IE-1 are the best defined and considered as immunodominant [31, 32]. Remarkably, this CMV-specific, fully functional effector memory pool of CD8+ T cells is able to persist for a long time, suggesting a unique differentiation pathway that apparently differs from the generation of short-lived CD8+ memory T cells generated after infection with acute viruses  or even closely-related persistent herpes viruses such as Epstein-Barr virus . The reasons for such differences are unknown but may be related to the type of cell hosting the latent virus.
It is assumed that the latent phase of CMV infection can be accompanied by recurrent episodes of viral reactivation, but this has been hard to document in humans . This is mostly a subclinical process, which leads to production of immunogenic transcripts that may maintain memory inflation of virus-specific cytotoxic lymphocytes. Supporting this idea it was demonstrated, in a murine model, that treatment with valaciclovir (that inhibits CMV replication) leads to a dramatic suppression of MCMV- specific T-cell memory subset and regeneration of the naive CD8+ T-cell compartment . This could be taken to represent a reversal of a classic marker of immunosenes- cence. Results from the mouse model also demonstrated that latent CMV infection impairs immunity in old animals . Such experiments have not been carried out on humans, but studies on people receiving anti-HSV drugs which also have some minor effects on T cell responses to CMV pp65 have been published . Results in mice are clear  and in humans recent data show that CMV viral load increases after the age of 70, suggesting potential slippage of the strict control of latency .
The total CMV-specific T-cell response in seropositive subjects constitutes on average approximately 10 % of both the CD4+ and CD8+ memory compartments, and can be far greater in older people. It has been documented that cross-reactive recognition of CMV proteins in seronegative individuals is rare and is limited exclusively to CD8+ T cells, so that most of the cells responding to CMV are likely to be truly CMV-specific and not cross-reactive . Because of the high level of both CD4+
and CD8+ memory T cells in older people, the idea arose that CMV promotes immunosenescence in a clinically-relevant sense, i.e. associated with mortality [30, 40-42]. This was based on original findings that an inverted CD4:8 cell ratio caused by accumulated CD8+ memory cells specific for CMV was associated with 2, 4 and 6-year mortality of very old people in the longitudinal Swedish OCTO/NONA studies. These studies remain some of the very few indicating detrimental effects of the immune changes observed, meeting the definition of “senescence” as being demonstrably detrimental. The cluster of parameters including lower numbers of B cells as well as T cell parameters, and CMV-seropositivity, was designated an “Immune Risk Profile” (IRP). However, due to lack of other studies of this type, it remains unclear whether the IRP is informative in other populations and whether it has any relevance to younger people not already selected for survival into advanced age. Different risk factors may be relevant in other populations, for example, in the BELFRAIL study of Flemings around 83 years old at baseline. In this study, it was found that a naive T-cell-dominated CD4:8 ratio >5 rather than the IRP-defining ratio of <1 in the Swedish population, and which was absent in young donors, was associated with a higher physical and functional impairment in very old people infected with CMV  . Moreover, further studies have revealed that this profile is associated with 3-year mortality in women, but not men (Adriaensen et al., unpublished observations). Unlike in the Swedish studies an IRP characterized by an inverted CD4:8 ratio was not informative for survival in this Belgian population. This finding suggests that in a very old population not only CMV infection but also the degree of concomitant immune dysregulation play an important role with regard to physical impairment and consequences for health state, and these may well be very sensitive to the early-life and current circumstances of the population studied. In fact, there may be disparities of such major proportions that in some cases at least it could be possible to discern an advantage to being CMV-infected in early life, as reflected in the “hygiene hypothesis”. For example, as mentioned above, there are some published data suggesting that that in young humans or young mice, CMV may improve immune responses to some antigens and to influenza virus, probably by way of increased pro-inflammatory responses [44, 45]. There is also an animal model in which CMV protects against fatal infections with certain other pathogenic organisms . These observations suggest that the effect of CMV on the immune system may be highly dependent also on an individuals’ age and circumstances, and that what is viewed as ageing is in fact later collateral damage from immune reactivity that was beneficial in earlier life [47, 48]. This is saying nothing more than that the same immune pathology that always accompanies immune responses to acute viruses is also caused by CMV, but over a chronic time scale and usually subclinical. Because the magnitude of the CMV-specific immune response increases with age and accordingly with the chronological history of co-existence with an antigen, this leads to ever-increasing memory inflation and potential collateral damage [30, 35, 40].
Another important variable that appears to drive higher expansions of memory CMV-specific effector T cells was shown to be a higher viral load , which, as mentioned above, is also greater in the aged . The influence of the viral inoculum on the degree of memory T-cell inflation may provide an explanation for the observed variations in the magnitude and phenotype of CMV-specific T-cell responses in people , only some of which is likely to have a genetic basis . Thus, as we saw above, in general the CMV-specific CD8+ T-cell response is characterized in the periphery by an accumulation of late-stage differentiated effector cells, sometimes referred to in the literature as “terminally” differentiated, which have downregulated expression of the costimulatory molecules CD27 and CD28. This population of cells is characterized by expression of the negative costimulatory molecules KLRG-1 and CD57 and re-expression of CD45RA. They have lost the homing molecules CCR7 and CD62L as well the cytokine receptors CD122 and CD127 [3,33,51,52]. Although some T cells with this phenotype are present in CMV-seronegative people, they are overwhelmingly present in CMV-seropositive adults . They maintain high levels of cytotoxic molecules such as granzyme B but many of them seem not to produce perforins and are therefore compromised in their killing ability. Inappropriate release of superfluous granzyme B has been postulated to cause much of the tissue damage associated with diseases of ageing . Many of these cells retain the ability to produce the pro-inflammatory cytokine IFN-y upon mitogenic stimulation but the fraction of those doing so upon stimulation with CMV antigens is lower in the elderly than in the young, at least for the one specific CMV pp65 epitope tested some years ago . In this respect at least, some of these accumulated late-stage differentiated CD8+ T cells may indeed be dysfunctional, potentially senescent, but this is not the case for all of them. Intriguingly, despite the loss of the expression of some costimulatory molecules and up-regulation of negative signaling receptors, it was demonstrated that polyfunctional CMV-specific CD8+ T cells were at an intermediate state of differentiation and were not necessarily restricted in their replicative capacity by excessive telomere erosion . As mentioned above, infection with CMV induces additionally the differentiation not only of CD4+ T cells, but also NK cells - all of which share effector characteristics of CMV-specific CD8+ T cells. It was supposed that this overlap in differentiation of these pools of lymphocytes might be regulated by shared transcriptional machinery ; thus, there could be a degree of interchangeability of immune elements maintaining control of CMV, and resulting in heterogeneous manifestations of the side-effects recognized as characteristic of immunosenescence.
It should not be forgotten that there is a second major type of T cells, those expressing уб antigen receptors rather than the more common ap receptors, which also display considerable age-related differences, mostly decreased in older adults. However, after infection with CMV, the usually numerically smaller pool of V62- negative уб T cells appears to be significantly increased and accumulates further throughout life (Fig. 4.2). In contrast, CMV-seropositivity has no influence on V62- positive уб T cells, which maintain a less differentiated phenotype . Moreover, the уб T-cell repertoire is more restricted in CMV-seropositive individuals. Taken together, these findings suggest that infection with CMV causes changes in уб T-cell repertoire similar to those described in CD8+ T cells [55, 56]. In a recent systems-l evel analysis of the immune system of healthy twins it was shown that CMV has indeed a major non-heritable impact on immune parameters (such as effector CD8+ and уб T cells) and can dramatically modulate the overall immune profile of healthy individuals .
It is conceivable that all these alterations might be important contributory factors to immunosenescence but certainly are not the whole story. Further functional and longitudinal studies are required to elucidate the relationship between ageing, CMV, immunosenescence and their clinical consequences [3, 41, 42]. However, a large body of mostly circumstantial and to a great extent controversial work has been published on the consequences of CMV infection for health, frailty and mortality.