Sci. Aging Knowl. Environ., 14 December 2005
Vol. 2005, Issue 50, p. pe39
[DOI: 10.1126/sageke.2005.50.pe39]


When T Cells Get Old

Graham Pawelec

The author is at the Center for Medical Research at the University of Tuebingen, D-72072 Tuebingen, Germany. E-mail: graham.pawelec{at}

Key Words: T cells • immunoscenescence • CMV • vaccination • remediation


Of all the age-associated changes to the immune system observed in humans and animal models, those affecting T cells appear to be the most grave (see "Immunity Challenge"). Age-associated changes to the distribution of T cell subsets (e.g., naïve T cells and various memory cell subpopulations) and their function (e.g., anergic and/or suppressive) may contribute to the reduced ability of elderly individuals to resist infectious diseases and to respond to vaccination. These alterations are globally referred to as "immunosenescence," a poorly defined state of clinically relevant decreased immunity. A workshop titled "Differentiation and Immunoscenescence in the Immune System" was held at the Edward Jenner Institute for Vaccine Research in Compton, UK, from 6 to 7 October 2005, to review the state of our knowledge in these areas and to formulate future avenues for investigation. The Jenner Institute was an appropriate location for the meeting, because the most profound clinical impact of age on the immune system concerns the response of the elderly to vaccination (see "All Pain, No Gain"). Organized by Peter C. Beverley of the Jenner Institute, Arne N. Akbar of University College, London, and Donald Palmer of the Royal Veterinary College, London, the meeting included discussions on T and B cell differentiation and aging, as well as dendritic cell and neutrophil data. This Perspective focuses on the main findings relevant to T cell immunosenescence, one of the most important areas of study for which most data are available.

T Cell Persistence and Homeostasis

Immunological memory--the ability of the immune system to respond more efficiently to a second encounter with the same pathogen--is defective in the elderly. Development of the total T cell pool in early life is dependent on the output of naïve T cells from the thymus and the establishment of an antigen-experienced pool of memory cells, which has a more restricted repertoire. The total number of T cells is more or less constant throughout life, but it appears that changes in the composition of each pool, determined by the balance of proliferation and death in T cell subpopulations, plays a role in the decline in immunocompetence. T cell subpopulations are classified by the presence of particular cell surface markers; for example, helper T cells carry the CD4 surface protein, and memory cells carry CD8. (For more information on the immune system, see "Immunity Challenge.") Diana L. Wallace and Peter C. Beverley, both at the Jenner Institute, in collaboration with Derek C. Macallan of St. George's Hospital Medical School in London, have been performing pioneering work on the kinetics of T cell turnover during human aging (1). By in vivo labeling of dividing lymphocytes with glucose that contains deuterium, the researchers were able to measure the rate at which deuterium is incorporated into DNA and then diluted with each cell division. Their studies have revealed that in healthy young people, cells with a memory/effector phenotype (e.g., CD45RO+ cells) divide more rapidly than naïve (e.g., CD45RA+) cells in both the CD4+ and CD8+ T cell subsets. CD45RO+ cells have a 26-day half-life, with 2.7% of cells dividing every day, whereas the RA+ cells have a 154-day half-life and only 0.5% divide every day. The rates at which CD4+ CD45RA+, CD4+ CD45RO+, and CD8+ CD45RO+ cells divide and disappear are the same among young and old, but the rate of disappearance of CD8+ CD45RA+ T cells was far slower in the elderly than in the young (by a factor > 10), implying that these cells represent a long-lived pool of cells that persists in the circulation. In addition, these persistent cells contained large clonal populations. Despite their RA+ phenotype, these are likely to be memory cells specific for some of the same epitopes identified by the other researchers, as discussed below.

These findings suggest that memory T cells turn over faster than naïve T cells, implying that replicative senescence will occur faster than in the naïve cell population. In addition, because CD8+ naïve cells are also turning over in the elderly, it is likely that each naïve cell of an elderly person has accumulated a larger number of population doublings during the life of the individual compared with a cell of a younger person. This observation provides an explanation for naïve cell aging, which is also supported by elegant work in mice, as presented by Laura Haynes of the Trudeau Institute at Saranac Lake, NY, showing that naïve phenotype cells in old mice are functionally compromised.

In the Wallace study, the CD8+ CD45RA+ T cells were associated with seropositivity for cytomegalovirus (CMV) in the young. It is possible that CMV in some way prevented cell death in virus-specific, or even bystander, CD8+ populations, which would explain why CD8+ CD45RA+ T cells persist. These data are consistent with many other studies presented at the conference, implicating CMV as a major driving force behind many of the measured manifestations of immunosenescence in humans (see also "T Cell Tunnel Vision" and Koch Perspective). Some of these findings are discussed below.

Role of CMV and Other Persistent Viruses

A large body of evidence suggests that CMV infection affects CD8+ cell turnover and function in both young and elderly individuals, but the effects of this virus on CD4+ cells are less well studied, and those on natural killer (NK) cells in the elderly are essentially unknown. However, preliminary studies by Paul A. Moss at the University of Birmingham, UK, indicate that CMV has relatively minor effects on NK cells, B cells, or regulatory T cells in the elderly. He showed that CMV negatively influences the number of naïve CD8+ cells in the elderly and suggested that CMV accelerates age-associated processes, which would occur regardless of viral infection. These observations are consistent with the much-discussed idea that chronic antigenic stimulation of any kind (including by CMV antigens) and at any age can drive clonal expansions of specific memory cell subpopulations, resulting in an overall deleterious effect on immune function. Clonal expansion of cell subpopulations decreases the available T cell repertoire, takes up space that could be used by newly developed cells, and gives rise to dysfunctional cells. Advanced age results in the accumulation of multiple such problems.

Along similar lines of thought, Janko Nikolich-Zugich of the Oregon Health and Science University in Portland presented strong evidence that antigen-independent, sustained proliferation of T cells over a lifetime, such as that induced by lymphopenia or by adjuvants, can induce persistent T cell clonal expansions in specific pathogen-free mice. Graham Pawelec of the University of Tuebingen pursued this line of investigation by comparing chronic antigenic stimulation caused by persistent activating viruses in the elderly with that caused by tumor antigens in mostly younger cancer patients. Many similarities in the CD8+ response were observed between the two, both functionally and phenotypically.

The accumulation of CMV-specific CD8+ cells bearing the hallmarks of anergic cells--dysfunctional cells that carry antigen-specific receptors but are unable to mediate their full range of responses upon antigen recognition--represents an important part of the immune risk phenotype (IRP) predictive of mortality in the very elderly. The IRP concept, which originally consisted of a cluster of relatively simple immunological parameters, such as the CD4+:CD8+ ratio, number of CD8+ CD28- T cells, and proliferative responses to mitogens, all of which are now believed to be influenced by CMV, emerged from the longitudinal aging studies carried out by Anders Wikby and colleagues at the University of Jönköping in Sweden (2). The IRP was further established through a long-term collaborative undertaking supported by the European Commission and coordinated by Pawelec (see the "T-Cells In Ageing" Web site for details and also the Koch Perspective) In addition to its use in establishing prognosis in older patients, the IRP may also prove useful for assessing "biomarkers of immunosenescence" in younger individuals. Recent work from Pawelec's group has shown that CD8+ clonal expansions, driven predominantly by CMV but also by Epstein-Barr virus, begin in early middle-age.

We now know that among the very elderly (e.g., >85 years), those individuals with a larger number of clonal expansions have a longer life expectancy than those with few clonal expansions (but note that all individuals possess very large numbers of CMV-specific CD8+ cells) (3). The model best explaining these data is one in which CMV drives clonal expansions of multiple CD8+ cells recognizing different CMV epitopes; in the young, up to 10% of the T cell repertoire is already obsessed with CMV, as was detailed in Moss's presentation. The number of different clones that are expanded increases with age, as does the overall number of CD8+ cells recognizing CMV epitopes. This clonal expansion leads to dysregulation of the immune response and is thus detrimental to individuals, but that may be the price to be paid to maintain a vitally necessary CMV immunosurveillance. However, at a very advanced age, clonal diversity starts to shrink, and individuals who possess only a small number of clonal expansions become at greater risk of death than those who retain a larger repertoire. Studies by Nikolich-Zugich in mice support the notion that virus infection drives clonal expansion. In those studies, persistent herpes simplex virus 1 infection led to the accumulation of clones specific for this virus and to their dysregulation over time, in the form of "memory inflation" (in other words, the persistent clonal expansions of memory cells). This inflation was, in turn, suppressed by continuous antiviral treatment of infected mice, showing directly that subclinical herpesvirus reactivation is necessary to drive the expansions of memory cells. It will be of interest to examine the clonal complexity and repertoire evolution in these animals.

As stated earlier, the role of CMV-specific CD4+ cells in this context has not yet been assessed. Arne N. Akbar, however, presented his studies on the functional status of CD4+ cells in the elderly. Unlike the studies on CD8+ cells, which revealed larger numbers of dysfunctional cells, the assays employed by Akbar's group only revealed the presence of nonanergic cells in elderly individuals. This is likely to be due to the fact that the use of major histocompatibility complex class II multimers, a method used for direct T cell receptor staining, remains fraught with technical difficulties. Nonetheless, his data on functional CMV-specific CD4+ cells were extremely informative. He reported that in older seropositive donors, CMV-specific cells had shorter telomeres, a sign of increased cell division, than cells from donors free of virus. Using CMV lysates, Akbar found that CMV stimulates the secretion of type I interferon (IFN-{alpha}) by dendritic cells, which, in turn, inhibits telomerase function in all virus-specific cells, not just the CMV- specific ones, and increases the fraction of CD27- CD28- cells, which are more differentiated and thus closer to senescence. These results may explain the unexpected finding that cytolytic T lymphocytes in CMV-seropositive individuals possess more differentiated phenotypes than those either CMV-specific or specific for other, unrelated antigens in seronegative donors.

Remediation Is Possible

The work discussed above suggests that targeted neutralization of CMV-stimulated IFN-{alpha} production might be beneficial in CMV-positive elderly individuals. Other possible therapies include the use of antiviral agents or therapeutic vaccination to eliminate CMV-specific T cell populations, but neither approach is likely to be available in the near future for use in the elderly, for several medical, ethical, financial, and logistical reasons. Similarly, adoptive immunotherapy with ex vivo-generated CMV-specific CD8+ cells, as applied in stem cell transplant recipients, is probably not ever going to be usable in the elderly. The elimination of the dysfunctional CMV-specific CD8+ cells might be beneficial in removing potentially suppressive cells or, at the very least, in making "space" for naïve cells that could, in theory, still be generated from residual thymic islands present in most elderly individuals. But how to target dysfunctional CMV-specific CD8+ cells, while leaving the functional ones intact, is a puzzle.

One possibility, discussed by Pawelec, might be to eliminate cells that are positive for both killer cell lectin-like receptor (KLR) G-1 and CD57 antigens, because it is only the small fraction of KLRG1+ but CD57- cells that seems to retain functionality in the elderly. This would be a way to eliminate dysfunctional CMV and T cell receptor (TCR)-positive CD8+ cells while keeping the functional ones. Another possibility might be to select CMV-specific cells with a naïve phenotype (a very small number of which appear to be present in most elderly individuals) and to expand them ex vivo using interleukin 15 (IL-15). As reported by Beverley, naïve CD8+ cells maintained in long-term culture in the presence of IL-15 retain their naïve phenotype, up-regulate telomerase activity, and actually increase the length of their telomeric repeats. This approach would obviously rely on the presence of naïve cells in the elderly, but work by Beatrix Grubeck-Loebenstein at the Institute for Biomedical Ageing Research in Innsbruck raised serious doubts about this possibility. She reported her odyssey in search of the "truly" naïve cell in the elderly. She began by isolating CD8+ CD28+ CD45RA+ CD62L+ cells. Theoretically, these cells should be naïve, but she found that they had shorter telomeres when isolated from elderly individuals as compared with young ones. In addition, clonal spectratyping revealed a smaller repertoire in the elderly. Siegfried Kohler and Andreas Thiel of the German Rheumatology Research Center in Berlin presented similar findings. They showed that CD4+ naïve-phenotype T cells can also be identified by their CD31+ phenotype. This subset of cells expresses a polyclonal TCR repertoire that is gradually lost with aging. In contrast, the CD31- subset has an oligoclonal repertoire that is stably maintained with aging. Although the precise history of this latter subset of cells remains to be defined, these results imply that even naïve cells in the elderly have "aged."

Similar results were obtained in mice, as described by Haynes and already mentioned above. She found that CD4+ naïve cells in old mice (which were CD28+, CD134+, and CXCR5+) showed decreased levels of expression of CD40L, a molecule important for activating dendritic cells and facilitating complete T cell stimulation. Concentrations of CD40L were increased by administering IL-2, the use of which might represent a potential avenue to treating immunoscenescence in humans. After all the CD4+ cells were depleted by antibody treatment and the animals were left to repopulate for 2 months, the newly formed naïve cells, even in the old animals, were perfectly functional. These results suggest that removing dysfunctional cells and allowing repopulation can indeed "rejuvenate" the T cell system. However, once again, it is unlikely that such an approach would ever be applied to elderly people.

On the other hand, attempting to increase the output of naïve cells from the thymus in elderly individuals, regardless of whether it is combined with peripheral cell depletion, may be a viable method for combating immunosenescence. To this end, Richard Aspinall of Imperial College, London, explained his strategy of specifically targeting IL-7 to the thymus, thereby avoiding general side effects due to cytokine administration. He engineered a fusion protein between IL-7 and the CCR9 ligand, the receptors for which are expressed exclusively by thymic stromal cells. On challenge with influenza, mice given the fusion protein had a decreased lung viral load and fewer tumor necrosis factor (TNF)-{alpha}-producing CD8+ cells. The success of this approach as a therapy for immunosenescence would require that T cell progenitors are as fit in the elderly as they are in the young, but according to the experiments reported by Ken Dorshkind at the University of California, Los Angeles, in the mouse this is not the case. He showed that "early T cell progenitors" (ETP; CD44+ CD25- c-kithigh CD127-) are fewer in number, are more prone to apoptosis, and have less proliferative capacity in old animals than in young ones. Thus, if there is a block in the potential of hematopoietic stem cells to generate ETP among elderly individuals, it is difficult to see how IL-7 supplementation could help.

One final therapeutic possibility, the old standby caloric restriction, was discussed by Nikolich-Zugich, not in the usual mouse model, but in the much more demanding and relevant rhesus monkey. Calorically restricted CMV+ monkeys were better able to maintain their numbers of naïve CD8+ cells and restrict the numbers of effector-memory cells than ad lib fed animals. They also had a greater number of TCR-excision circle-bearing cells (in other words, cells with less of a proliferative history) and lower concentrations of proinflammatory factors (such as TNF-{alpha} and IFN-{gamma}). Thus, even when considering the deleterious effects of CMV infection, the advice to eat less might still be the best available (see Masoro Review).


The presentations highlighted in this Perspective show that CMV infection has profound effects on the aging immune system, particularly in driving the common clonal expansions of CD8+ cells, which is observed beginning in early middle age and results in the emergence of increasing numbers of different clones of T cells specific for CMV. This, in turn, leads to the eventual shrinking of the T cell repertoire at a very old age, which is associated with incipient mortality. Many questions, however, remain. If CMV is really having such a disastrous effect in "accelerating" immunosenescence in the elderly, is it doing the same thing in the young? Are CMV- elderly donors more healthy that their CMV+ counterparts, and do they live longer? (Are centenarians CMV- seronegative?) Is it indeed the case that CD4+ and CD8+ cells are equally affected, but B cells, dendritic cells, and NK cells are not? Are all types of T cells equally affected (in other words, what about T regulatory cells)? Does CMV reactivate more frequently in elderly individuals at risk for immunoscenescence, or does the duration of infection determine the level of immunosenescence? Subsequent studies from some of the scientists who presented their work at the Jenner Institute this year may be able to answer some of these questions.

December 14, 2005
  1. D. L. Wallace, Y. Zhang, H. Ghattas, A. Worth, A. Irvine, A. R. Bennett, G. E. Griffin, P. C. Beverley, D. F. Tough, D. C. Macallan, Direct measurement of T cell subset kinetics in vivo in elderly men and women. J. Immunol. 173, 1787-1794 (2004).[Abstract/Free Full Text]
  2. A. Wikby, F. Ferguson, R. Forsey, J. Thompson, J. Strindhall, S. Lofgren, B. O. Nilsson, J. Ernerudh, G. Pawelec, B. Johansson, An immune risk phenotype, cognitive impairment, and survival in very late life: Impact of allostatic load in Swedish octogenarian and nonagenarian humans. J. Gerontol. A Biol. Sci. Med. Sci. 60, 556-565 (2005).[Abstract/Free Full Text]
  3. S. R. Hadrup, J. Strindhall, T. Køllgaard, T. Seremet, B. Johansson, G. Pawelec, P. Straten, A. Wikby, Longitudinal studies of clonally expanded CD8 T cells reveal a repertoire shrinkage predicting mortality and an increased number of dysfunctional CMV-specific cells in the very elderly. J. Immunol., in press.
Citation: G. Pawelec, When T Cells Get Old. Sci. Aging Knowl. Environ. 2005 (50), pe39 (2005).

Science of Aging Knowledge Environment. ISSN 1539-6150